ArticlePDF Available

Abstract and Figures

A bombardment of information overloads our sensory, perceptual and cognitive systems, which must integrate new information with memory of past scenes and events. Mechanisms employed to overcome sensory system bottlenecks include selective attention, Gestalt gist perception, categorization, and the recently investigated ensemble encoding of set summary statistics. We explore compensatory cognitive processes focusing on categorization and set ensemble summary statistics that relate objects sharing properties or function. Without encoding individual details of all individuals, (or as a shortcut to representing these details), observers perceive category prototype and boundaries or set mean and range, and perhaps higher order statistics as well. We found that observers perceive set mean and range, automatically, implicitly, and on-the-fly, for each presented set sequence, independently, and we found matching properties for category representation, suggesting a similar computational mechanism underlies the two phenomena. But categorization depends on a lifetime of learning about categories and their prototypes and boundaries. We now developed novel abstract “amoeba” forms, which are complex images similar to categories, but have simple ultrametric structure that observers can categorize on-the-fly (rather than depending on pre-learned categories). We find that, not only do observers learn the amoeba categories on-the-fly, they also build representations of their progenitor (related, but not equivalent, to set “mean” or category prototype), as well as category boundaries (related to set range and inter-category boundaries). These findings put set perception in a new light, related to object, scene and category representation.
Content may be subject to copyright.
1 23
Attention, Perception, &
Psychophysics
ISSN 1943-3921
Volume 81
Number 8
Atten Percept Psychophys (2019)
81:2850-2872
DOI 10.3758/s13414-019-01792-7
Relating categorization to set summary
statistics perception
Noam Khayat & Shaul Hochstein
1 23
Your article is published under the Creative
Commons Attribution license which allows
users to read, copy, distribute and make
derivative works, as long as the author of
the original work is cited. You may self-
archive this article on your own website, an
institutional repository or funder’s repository
and make it publicly available immediately.
Relating categorization to set summary statistics perception
Noam Khayat
1
&Shaul Hochstein
1
#The Author(s) 2019
Abstract
Two cognitive processes have been explored that compensate for the limited information that can be perceived and remembered
at any given moment. The first parsimonious cognitive process is object categorization. We naturally relate objects to their
category, assume they share relevant category properties, often disregarding irrelevant characteristics. Another scene organizing
mechanism is representing aspects of the visual world in terms of summary statistics. Spreading attention over a group of objects
with some similarity, one perceives an ensemble representation of the group. Without encoding detailed information of individ-
uals, observers process summary data concerning the group, including set mean for various features (from circle size to face
expression). Just as categorization may include/depend on prototype and intercategory boundaries, so set perception includes
property mean and range. We now explore common features of these processes. We previously investigated summary perception
of low-level features with a rapid serial visual presentation (RSVP) paradigm and found that participants perceive both the mean
and range extremes of stimulus sets, automatically, implicitly, and on-the-fly, for each RSVP sequence, independently. We now
use the same experimental paradigm to test category representation of high-level objects. We find participants perceive categor-
ical characteristics better than they code individual elements. We relate category prototype to set mean and same/different
category to in/out-of-range elements, defining a direct parallel between low-level set perception and high-level categorization.
The implicit effects of mean or prototype and set or category boundaries are very similar. We suggest that object categorization
may share perceptual-computational mechanisms with set summary statistics perception.
Keywords Categorization .Prototype .Boundary .Summary statistics .Ensemble .Mean .Range
Categorization is one of the most important mechanisms for
facilitating perception and cognition, helping to overcome
cognitive-perceptual bottlenecks (Cowan, 2001;Luck&
Vo g e l , 1997) and perceive the gistof the scene (Alvarez &
Oliva, 2009; Cohen, Dennet & Kanwisher, 2016;Hochstein&
Ahissar, 2002; Hock, Gordon, & Whitehurst, 1974;Iordan,
Greene, Beck, & Fei-Fei, 2015,2016; Jackson-Nielsen,
Cohen & Pitts, 2017; Oliva & Torralba, 2006;Posner&
Keele, 1970). Categorization follows and expands on the nat-
ural categories of objects in our environment, the intrinsic
correlational structure of the world (Goldstone &
Hendrickson, 2010; Rosch, Mervis, Gray, Johnson, &
Boyes-Braem, 1976). There is long-term debate concerning
the mechanisms and cerebral sites of categorization, with re-
cent studies suggesting that there are multiple sites and pro-
cesses of categorization (Ashby & Valentin, 2017;Nosofsky,
Sanders, Gerdom, Douglas, & McDaniel, 2017). Thus, cate-
gorization itself may be categorized by task or goal (Ashby &
Maddox, 2011), neural circuit (Iordan et al., 2015; Nomura &
Reber, 2008), utility (J. D. Smith, 2014), and context
(Barsalou, 1987;Koriat&Sorka,2015,2017;Roth&
Shoben, 1983). The most common and accepted theoretical
mechanisms for categorization are still rule based, defining
clear boundaries between categories (Davis & Love, 2010;
Goldstone & Kersten, 2003; Sloutsky, 2003; E. E. Smith,
Langston, & Nisbett, 1992) and their cortical representations
(Iordan et al., 2015,2016; Kriegeskorte et al., 2008), and
prototype-based or exemplar-based, defining family resem-
blance (Ashby & Maddox, 2011; Goldstone & Kersten,
2003;Iordanetal.,2016; Maddox & Ashby, 1993;Medin,
Altom, & Murphy, 1984;Nosofsky,2011; Posner and Keele,
1968;Rosch,1973; Rosch, Mervis, et al., 1976; see also
Clapper, 2017).
In parallel, recent interest has focused on the perception of
summary statistics of sets of stimulus elements. Observers
have a reliable representation of the mean and range of sets
of stimuli, even without reliable perception of the individual
members of the presented set. Summary statistics, rapidly
*Shaul Hochstein
shaulhochstein@gmail.com
1
Life Sciences Institute and Edmond and Lily Safra Center (ELSC) for
Brain Research, Hebrew University, 91904 Jerusalem, Israel
https://doi.org/10.3758/s13414-019-01792-7
Attention, Perception, & Psychophysics (2019) 81:28502872
Published online: 26 June 2019
extracted from sets of similar items, presented spatially
(Alvarez & Oliva, 2009;Ariely,2001)ortemporally
(Corbett & Oriet, 2011; Gorea, Belkoura, & Solomon, 2014;
Hubert-Wallander & Boynton, 2015), include average, and
range or variance of their size (Allik, Toom, Raidvee,
Averin, & Kreegipuu, 2014;Ariely,2001;Corbett&Oriet,
2011; Morgan, Chubb, & Solomon, 2008;Solomon,2010),
orientation (Alvarez & Oliva, 2009), brightness (Bauer,
2009), spatial position (Alvarez & Oliva, 2008), and speed
and direction of motion (Sweeny, Haroz, & Whitney, 2013).
Extraction of summary statistics appears to be a general mech-
anism operating on various stimulus attributes, including low-
level information, as mentioned above, and more complex
characteristics, such as facial expression (emotion) and gender
(Haberman & Whitney, 2007,2009; Neumann,
Schweinberger, & Burton, 2013), object lifelikeness
(Yamanashi-Leib, Kosovicheva, & Whitney, 2016), biological
motion of human crowds (Sweeny, Haroz, & Whitney, 2013),
and even numerical averaging (Brezis, Bronfman, & Usher,
2015; for recent reviews, see Bauer, 2015; Cohen et al., 2016;
Haberman & Whitney, 2012; Hochstein, Pavlovskaya,
Bonneh, & Soroker, 2015). Examples of the methods used
in these studies are shown in Fig. 1; see methodological details
in the figure caption.
We have suggested that these phenomena, categorization
and set perception, may be related since they share basic char-
acteristics (Hochstein 2016a,2016b; Hochstein, Khayat,
Pavlovskaya, Bonneh, & Soroker, 2018). In both cases, when
viewing somewhat similar, but certainly not identical items,
we consider them as if they were the same, as a shortcut to
representing them and prescribing a single appropriate re-
sponse (Ariely, 2001; Medin, 1989; Rosch & Mervis, 1975;
Rosch, Mervis, et al., 1976). When we globally spread atten-
tion, and see a flock of sheep in a meadow, a shelf of alcohol
bottles at a bar, a line of cars in traffic, or a copse of trees in a
forest, we are both categorizing these objects as sheep, alcohol
bottles, cars, and trees, and relating to the average properties
of each set. Similarly, in laboratory experiments, we present a
set of circles (Alvarez & Oliva, 2008; Ariely, 2001; Corbett &
Oriet, 2011; Khayat & Hochstein, 2018), line segments
(Khayat & Hochstein, 2018; Robitaille & Harris, 2011), or
faces (Haberman & Whitney, 2007,2009), and observers per-
ceive the nature of the images as circles, lines, or faces and
relate to their average properties. All animals in the category
dogshave four legs and a tail, but they may vary in color,
size, and so forth. All circles in a set are round, though they
may vary in size or brightness. Categorization emphasizes
relevant or common properties and deemphasizes irrelevant
or uncommon properties, reducing differences among catego-
ry members (Fabre-Thorpe, 2011; Goldstone & Hendrickson,
2010; Hammer, Diesendruck, Weinshall, & Hochstein, 2009;
Rosch, Mervis, et al., 1976; Rosch, Mervis, et al., 1999, Rosch
&Lloyd,1978,Rosch2002). Similarly, set perception
captures summary statistics without noting individual values.
Categorization, like ensemble perception, may depend on rap-
id feature extraction, to determine presence of defining char-
acteristics of objects.
In particular, set perception includes set mean and range
(Ariely, 2001; Chong & Treisman, 2003,2005; Khayat &
Hochstein, 2018; Hochstein et al., 2018), and categorization
might rely on the related properties of prototype (or mean
exemplar; e.g. Ashby & Maddox, 2011) and/or intercategory
boundaries (or category range; e.g., Goldstone & Kersten,
2003). This conceptual similarity has been confirmed by the
recent finding that set characteristics are perceived implicitly
and automatically (Khayat & Hochstein, 2018), just as objects
are categorized implicitly and automatically at their basic cat-
egory level (Potter & Hagmann, 2015; Rosch, Mervis, et al.,
1976). Finally, it has been suggested that determining whether
a group of objects in a scene belong to the same category may
actually depend on their characteristics that allow them to be
seen as a set (Utochkin, 2015). The similarities of categories
and sets led us to ask if the detailed properties of their percep-
tion are also similar, so that it may be hypothesized that similar
mechanisms are responsible for their cerebral representation.
The goal of the current research is to detail the similarity
between set and category perception by applying to categories
the very same tests that we used to study implicit set percep-
tion (Khayat & Hochstein, 2018). The following section brief-
ly reviews the results of these previous tests.
We note in advance that there are important differences
between categorization and set perception. Object categories
are learned over a lifetime of experience, while set ensemble
statistics can be acquired on the fly. Different life experience
may lead to individual differences in categorization and
choice of object seen as the category prototype.
Categorization may involve semantic processes, while set per-
ception has been demonstrated for simple visual features
(though including face emotion). Thus, it would be difficult
to claim that ensemble perception and categorization are iden-
tical, or take place at the same cortical site. However, their
being different makes comparing them even more important,
since if they share essential properties, they may depend on
similar or analogous processes, albeit at different cortical sites.
This is the aim of the current study.
Previous study
We studied implicit perception and memory of set statistics by
presenting a rapid serial visual presentation (RSVP) sequence
of images of items differing by low-level properties (circles of
different size, lines of different orientation, discs of different
brightness; see Fig. 1b), and testing only memory of the seen
members of the sequence (Khayat & Hochstein, 2018). Note
that the mean of the setthe mean size circle, mean
Atten Percept Psychophys (2019) 81:28502872 2851
orientation line, or mean brightness diskwas sometimes in-
cluded in the set sequence and sometimes not. Following set
RSVP presentation, we presented two images simultaneously,
side by side. One of these images was of an item that had been
seen in the image sequencethe SEEN itemand one was a
NEW item, not seen in the sequence. Observer memory was
tested by asking participants to choose which of the two si-
multaneously presented image items had been seen in the
sequence. Participants were informed that always one item
had been SEEN and one would be NEW. We did not inform
them that sometimes one test element would have the property
that was the mean of all of the items presented in the sequence
and that this test item could be the SEEN item (i.e., a member
of the RSVP sequence, in which case it was, of course, includ-
ed in the sequence) or it could be the NEW item, (i.e., not a
member of the previously viewed sequence, and thus, in this
case, it had not been presented in the RSVP sequence). We
also did not inform them that sometimes the NEW, non-
sequence-member was outside the range of the properties of
the seen sequence elements. Not mentioning to the partici-
pants the words meanand range,the goal was to test
whether observers would automatically perceive set mean
property and choose the test item that matched this mean
irrespective of whether this test item was the one that had been
seen in the sequence or if it was the foil, the test item that was
new and never been seen before. Similarly, would observers
automatically perceive the range of the properties of the set
and easily reject foils that were outside the range of the items
in the sequence?
We call these test-stimulus contingencies trial subtypes, as
shown in Table 1.Weindicateasintest elements within the
range of the variable property of the sequence; outindicates
an element with this property outside that range, and mean
indicates a test element with test property equal to the mean of
all those in the sequence (note that to be the mean, the element
must be inthe sequence property range). Test stimuli consist
of a pair of images, one SEEN and one NEW, and we indicate
the pair with two mnemonics: the first mnemonic refers to the
test element SEEN in the sequence; the second to the NEW,
never-before-seen element, as follows: SEENmeanNEWin
(test element that was SEEN in the sequence equals the set
mean; both elements in the range of the variable property in
a
b
c
d
Fig. 1 Previous study stimulus sets. a Arielys(2001) schematic repre-
sentation of the two intervals used in his experiments trials. Observers
were exposed for 500 ms to a set of spatially dispersed circlesdiffering by
size and then asked if a test stimulus size had been present in the set, or, is
smaller/larger than the set mean. bKhayat and Hochsteins(2018)RSVP
sequences consisted of 12 elements, each presented for 100 ms plus
100 ms interstimulus interval (ISI), followed by a two-alternative
forced-choice (2-AFC) membership test (i.e., which test element had been
present in the sequence). Blocks contained circles differing in size, lines
differing in orientation, or discs differing in brightness. Observers were
asked which of two test elements was present in the set. They were
unaware that either test element could equal the set mean orthe nonmem-
ber could be outside the set range. cHaberman and Whitneys(2009)task
included four faces (from a set of 4, 8, 12, or 16), differing in facial
emotional expression, presented for 2 s. Observers then indicated whether
the test face was a member of the set, or was happier/sadder than the set
mean. dBrezis et al.s(2015) trials consisted of two-digit numbers se-
quentially presented in a rate of 500 ms/stimulus. Set size was 4, 8, or 16.
Participants reported their estimate of the set average
Atten Percept Psychophys (2019) 81:28502872
2852
the sequence); SEENinNEWmean (the property of the
never-seen NEW test element equals the mean of the seen
sequence elements); SEENinNEWin(bothtestelements
have the property within the range of the sequence elements,
but neither equals their mean); SEENmean-NEWout and
SEENin-NEWout (the property of the NEW, never-seen ele-
ment is outside the set range, and the property of the SEEN
test element is either equal to the mean or just in the sequence
range).
As demonstrated in Fig. 2ad, we found a mean effect for
each of the three variables tested, circle size, line orientation,
and disk brightness: Participants chose the test element with
the property that was equal to the mean more often, whether it
was the SEEN element (SEENmeanNEWin), or the NEW
element (SEENinNEWmean), compared with the case where
both were in the sequence range, but neither was the mean
(SEENinNEWin).
We concluded that since the stimulus sequence was quite
rapid, participants had difficulty remembering all the members
of the RSVP set, and maybe even any one of them. Instead,
they automatically used their implicit perception of the se-
quence set mean and range to respond positively to test ele-
ments that matched or were close to the set mean. Thus, per-
formance was more accurate for test SEEN elements that
equaled the meanSEENmeanNEWin (see Fig. 2ab, middle
bars; Fig. 2cd, left bars). When the NEW test element was
equal to the set mean, it was frequently chosen as if it were a
member (i.e., as if it had been seen in the set sequence).
Participants actually chose this mean NEW element more fre-
quently than the actual nonmean SEEN elementSEENin
NEWmean (see Fig. 2ab, leftmost bars; note that accuracy
below 0.5 means that the NEW element was chosen more
frequently than the SEEN one.)
In addition, we found a range effect (i.e., participants
rejected out-of-range nonmembers; SEENmeanNEWout
and SEENinNEWout) more frequently than in-range NEW
test elements (SEENmeanNEWin, SEENinNEWin,
SEENinNEWmean). This is shown in Fig. 2ab, right two
bars, and in Fig. 2ef, right bars, compared with left bars in
each graph. The same effect was seen for response time (RT;
Fig. 2g), which was shorter for out-of-range than in-range
NEW test elements, indicating they were rejected more rapid-
ly as well as more frequently.
We concluded that participants automatically and implicit-
ly determined the mean and range of the RSVP sequence even
though they were not instructed to do so and even though this
had no bearing on performance of the task at hand, which was
just to try to remember the seen sequence elements.
Furthermore, they did so on the fly for each trial, independent-
ly, since each trial had a different sequence mean and range.
Perception of set mean and range is not only implicit. In
another study, Hochstein et al. (2018) asked observers to ex-
plicitly compare means of two arrays of variously oriented
bars (mean comparison) or report presence of a bar with an
outlier orientation among the array elements (outlier detec-
tion). It was found that mean comparison depended on the
difference between the array means, and outlier detection
depended on the distance of the target from the array range
edge (see also Hochstein, 2016a,2016b; Hochstein,
Khayat, Pavlovskaya, Bonneh, & Soroker, 2018). Thus, both
set mean and range are perceived both explicitly and
implicitly.
The goal of the current study is to test whether there are
identical effects in the related perceptual phenomenon of
categorization.
Experiment 1. Category prototype
and boundary effects
Prototypes as averages
We investigate here the nontrivial comparison between stim-
ulus sets and object categories. The stimuli in previous studies
Table 1 Member recall test trial subtypes
SEEN test image (correct) NEW test image (Incorrect) Expected performance
SEENmean NEWout Best
SEENin NEWout Better
SEENmean NEWin Better
SEENin NEWin Baseline
SEENin NEWmean Worse
Note. Test image elements could be both from the RSVP sequence (in), one could be the mean (mean) of that sequence (whether presented,
SEENmean, or not, NEWmean), and the NEW element image could be out of the sequence range (NEWout). On every trial, one element image had been
SEEN in the sequence, and the other was not (i.e., NEW). Test pairs of the baseline subtype have both SEEN andNEW objects from the sequence range,
one actually present and one not, and neither is the mean. If participants have difficulty recalling all elements in the sequence, but perceive and recall the
mean of the sequence, we expect better performance when the SEEN test element equals the mean, and worse performance when the NEW element
equals the mean. If participants perceive the range of the sequence elements, we expect better performance when the NEWelement is outside the range
and easily rejected. Trial subtypes were presented in randomized order, without observers knowing about this classification
Atten Percept Psychophys (2019) 81:28502872 2853
of statistical perception were very similar, in each case, usu-
ally differing by a single varying feature (e.g., Ariely, 2001;
Corbett & Oriet, 2011), or a combination of features forming a
single high-level feature (e.g., facial expression; Haberman &
Whitney, 2007,2009). In contrast, categories might be
thought of as a set of objects composed of combinations of
multiple features, with only some of these features necessarily
present in each category exemplar (where membership is de-
fined by family resemblance). Thus, we compare the mean of
the set elements with the prototype of category exemplars,
based on the view that prototypes are the central or most
common representations of a category (Goldstone &
Kersten, 2003), possessing the mean values of its attributes
(Langlois & Roggman, 1990;Reed,1972; Rosch & Lloyd,
0.4
0.5
0.6
0.7
0.8
0.9
1
SEENin
NEWmean
SEENin
NEWin
SEENmean
NEWin
SEENin
NEWout
SEENmean
NEWout
Proporon correct
Accuracy by subtype & feature
Size
Orientaon
Brightness
a
SEENin
NEWmean
SEENin
NEWin
SEENmean
NEWin
SEENin
NEWout
SEENmean
NEWout
Accuracy by subtype all features
b
Low-level experiment results
0.4
0.5
0.6
0.7
0.8
0.9
1
mean not mean
Proporon correct
Mean effect by feature
Size
Orientaon
Brightness
c
SEEN test element
mean non-mean
SEEN test element
Mean effect all features
***
d
0.4
0.5
0.6
0.7
0.8
0.9
1
in range out of range
Proporon correct
Range effect by feature
Size
Orientaon
Brightness
e
in range out of range
f
***
Range effect all features
800
900
1000
1100
1200
1300
in range out of range
Time (ms)
Response Time
***
g
NEW test element
Trial subtype test elements (NEW - SEEN)
***
***
Fig. 2 Low-level experiment results. aAccuracy rates for each trial
subtype (i.e., their test elements); SEEN versus NEW being equal to the
set sequence mean (mean), being in the set range (in) or outside the
range (out), and each stimulus feature (colored bars; see legend). Thus,
trial subtypes include: SEENmeanNEWin (seen test element = mean;
both test elements in sequence range); SEENinNEWmean (new test
element = mean; both in sequence range); SEENinNEWin (neither =
mean; both in sequence range); SEENmeanNEWout (seen test element
= mean; new element outside sequence range); SEENinNEWout (seen
test element not = mean; new test element outside sequence range). b
Accuracy rates for each trial subtype, averaged across stimulus features. c
Mean effect for each stimulus feature; accuracy rates for trials where the
SEEN test element equaled the set mean versus when it differed from the
mean. Each comparison is significant, p< .05. dMean effect across
features, p<.001.eRange effect for each stimulus feature; accuracy rates
for trials where the NEW testelement is in range versus out of range. Each
comparison is significant, p<.01.fRange effect across features, p<.001.
gRange effect seen in response time, indicating this is not an accuracy
time trade-off, p< .001. All results from Khayat and Hochstein (2018).
Error bars here and in all following graphs represent between-participant
standard error of the mean. (Color figure online)
Atten Percept Psychophys (2019) 81:28502872
2854
1978; Rosch, Mervis, et al., 1976; Rosch, Simpson, & Miller,
1976). Note, however, that comparing these perceptual proce-
dures does not depend on this definition of prototype, or even
on prototype theory itself. Comparing categorization with set
summary perception is valid simply because in both cases
several stimuli are perceived as belonging together, perhaps
inducing the same response, because they share some charac-
teristics and differ in others.
Similarly, we compare knowledge of category boundaries
with perception of set range edges. As shown above, perceiv-
ing set range edges allows for rapid detection of outlier ele-
ments, and even unconscious perception of these edges allows
for rapid rejection of out-of-range elements when trying to
remember which elements were previously viewed. This
was called the range effect(Khayat & Hochstein 2018).
Similarly, knowing category boundaries allows for rapid sep-
aration of objects that belong to different categories, which we
shall call a boundary effect.Thus, we compare properties of
set perception and categorization in terms of observersim-
plicit determination and knowledge of both the set mean and
category prototype, as well as, the set range edges and the
category boundaries. That is, having found that observers per-
ceive rapidly and implicitly the mean and range of element
sets, and that they use this information when judging memory
of sequence stimuli, we now test if the same characteristics are
present for object categories. Do observers of a sequence of
objects determine automatically and implicitly their category
and use the implied prototype (whether shown or not shown in
the sequence) and the boundaries of the implied category,
when later choosing images as having been seen in the se-
quence? These will be called the prototype and boundary ef-
fects, respectively. If we find similar characteristics in these
processes, for categorization as for set perception, we will
suggest that they may share basic perceptual-cognitive
mechanisms.
We note at the outset that there are important differences
between perceiving set summary statistics and categorizing ob-
jects. We perceive the mean size, orientation, brightness, and so
forth, of sets that we see just once, sets which are unrelated to
any other sets seen before. Presented with a set of images,
sequentially or simultaneously, we derive the mean and range
of the size, orientation, brightness, and so forth, of that set, on
the fly and trial by trial. Thus, presented with a single stimulus
in isolation, it is logically inconsistent to ask to what set it
belongs. In contrast, by their very nature, categories are learned
over a lifetime of experience, and with this knowledge, we can
know immediately to what category a group of objects, or even
a single object belongs. In fact, one of the defining characteris-
tics of basiccategories is that these are the names given to
single objects (e.g., cat, car, fork, apple; Potter & Hagmann,
2015; Rosch, Mervis, et al., 1976). The situation with catego-
rization is unlike that with sets, where we derive the set mean,
on the fly, as we are presented with set members. Instead, when
encountering an object (or group of objects belonging to a
single category), we know the category to which it belongs,
and we also know what is the prototype of that category and
the category boundaries; there is no need, and no possibility, of
deriving anew the category, prototype, and boundaries of a
group of familiar objects (though we can learn new categories
of unfamiliar objects; see Hochstein et al., 2019). Furthermore,
categories may be learned and recognized semantically, while
the basic features of sets are often nonsemantic. Nevertheless,
and this is the basic argument of the current study, there may be
similarities, if not identities, of mechanisms for representing set
means/ranges and category prototypes/boundaries. We set out
here to find the degree of similarity between these very different
phenomena before endeavoring to uncover underlying mecha-
nisms. Finding similarities, despite the differences enumerated
above, would suggest that there are relationships between low-
level and high-level representations of images, objects, catego-
ries, and concepts.
Method
We present rapid stimulus visual presentation (RSVP) se-
quencesofimagesofhigh-levelcategoryobjects,conditions
known to impair focused attention to each stimulus, but main-
tain statistical and categorical representations across time
(Corbett & Oriet, 2011; Potter, Wyble, Hagmann, &
McCourt, 2014). We then present two images, one identical
to one of the images in the sequence (the SEENimage) and the
other an image of a novel object (the NEW image). Observer
task is to choose the SEEN imagethe image that waspresent
in the sequence. This is a two-alternative forced-choice (2-
AFC) test, which is thus criterion free, and has a chance guess-
ing level of 50% (see Fig. 3).
We do not inform observers that one of the imaged objects
(either the SEEN or the NEW object) may be prototypical of
the sequence category, and one (the NEW object) may be
outside the sequence category (i.e., belong to another catego-
ry). Note that when NEWobjects were chosen from a different
category, still, they were purposely chosen to be not too distant
from the sequence categorythat is, from a relatively close
category (i.e., for basic level categories, a NEW object from
the same superordinate category; all NEW objects from the
same biological, nonbiological, or abstract concept groups of
Tab le 2; for example, for the category mammal, a nonmammal
animal; for dogs, a different mammal; for trees, another plant;
for food, a drink; for weapon, a screwdriver; for toy, a sand
clock). We hypothesize that the influence of prototypes on
implicit categorization and thus on memory will be similar
to the influence of the mean when we tested set item memory
(Khayat & Hochstein, 2018). Thus, we expect observers to
accept prototypical objects as SEEN more frequently (irrespec-
tive of whether they were in the sequence). Additionally, the
Atten Percept Psychophys (2019) 81:28502872 2855
presence in the test pair of an object outside the sequence cat-
egory may aid in rejecting it as not seen in the sequence, and
thus, NEW, just as items outside the set range were more easily
rejected as NEW and not SEEN (see Fig. 2ef).
Participants
Data of 15 in-house participants, students at the Hebrew
University of Jerusalem, were included in the analysis of
Experiment 1 (age range = 2027 years, mean = 23.4 years;
four males, 11 females). We also have results for 226 Amazon
Mechanical Turk (MTurk) participants for Experiment 3.
Participants provided informed consent and received compen-
sation for participation and reported normal or corrected-to-
normal vision.
Stimuli and procedure
Procedures for Experiment 1 took place in a dimly lit room,
with participants seated 50 cm from a 24-in. Dell LCD mon-
itor. We have less information as to their identity and precise
experimental conditions of the Experiment 3 Amazon MTurks
(we excluded ~25% of these data for trials with RTs <200 ms
or >4 s and for subjects with <33% remaining trials or <60%
correct responses overall, thus including as many trials/
subjects as possible, excluding data that are clearly not re-
sponses to the stimulus; e.g., Fabre-Thorpe, 2011). Stimuli
were generated using Psychtoolbox Version 3 for MATLAB
2015a (Brainard, 1997). MTurk testing used Adobe flash.
Images, chosen from the Google Images database, were pre-
sented against a gray background (RGB: 0.5, 0.5, 0.5).
Stimuli consisted of rapid serial visual presentation (RSVP)
of a sequence of high-level objects or scene images presented
in the center of the display, with a fixed size of 10.4-cm high ×
14.7-cm wide, as demonstrated in Fig. 3(see also examples of
images in Fig. 8). Experiment 1 was divided into three blocks
of 65 RSVP trials each, with a short break between them, to
complete 195 trials total per participant; Experiment 3 had 60
trials total for MTurk observers; one session/participant.
A set of images (12 for in-house students; nine for MTurks)
was presented in each RSVP sequence, with 167 ms stimulus
onset asynchrony (100 ms stimulus + 67 ms interstimulus
interval), and the sequence was followed by a 100 ms masking
stimulus. Then, after 1.5 s, two images were presented side by
side, simultaneously, for the membership test; one, an object
image that was SEEN in the sequence, and one a novel, NEW
object image. Sequence SEEN and NEW images were ran-
domly placed to the left and right of fixation in the middle half
of the width and height of the screen, and participants indicat-
ed position of the SEEN image by key press. Images remained
present until observer response. Since participants tend to per-
ceive and remember better early and late elements, known as
primacy and recency effects, in general and specifically in
summary representations (Hubert-Wallander & Boynton,
Fig. 3 High-level category RSVP membership tests. Example RSVP trial
with mammals as the set category. On the membership test, one of the
optional subtype pair of images (see Table 3) was presented for the SEEN
and the NEW images. The five trial subtypes for each of the 39 categories
are designed by choice of the test images.A SEEN object image could be
either SEENin (regular object image that was seen in the sequence and is
a member of the category, not the prototype) or SEENprot (seen in the
sequence and a prototype of the set category), while the nonmember
object could be NEWin or NEWprot (object image from the same cate-
gory but not included in the sequence, or the category prototype, again not
presented in the sequence), and could also be NEWout (belong to a dif-
ferent category)
Atten Percept Psychophys (2019) 81:28502872
2856
2015), we excluded from the test member images the first and
last two RSVP sequence images.
Thirty-nine categories (20 for MTurks of Experiment 3)
were included in the experiment (see Table 2), including
manmade and natural objects (animate, inanimate, and plants),
and abstract conceptual scenes from different category levels.
Each category was repeated in each trial subtype (see below),
with entirely different images for each trial. For each category,
we chose the three images that seemed to us to be closest to
prototypical, and used them in the three test subtypes
Table 2 Categories with examples of their prototypes and other exemplars
Category level Exemplar types
Superordinate level Basic level Typical exemplars (Prototypes & Common) Nonprototype exemplar
Plants* Potted plant, Cactus Watermelon plants, Vine
Trees* Oak, Olive Sequoia, Baobab
Fruits* Apple, Orange Pomegranate, Litchi
Animals* Dog, Deer Mosquito, Octopus
Reptiles Python, Iguana Legless Lizard, Commodore
Birds* Owl, Pigeon Penguin, Pelican
Mammals* Cow, Lion Whale, Bat
Dogs* German Shepard, Labrador Chi Wawa, Bull-Terrier
Food* Pasta, Pancakes Cake, Sushi
Weapons* Pistol, Riffle Cannon, Molotov bottle
Books Harry Potter, The Bible The Hobbit, Comics
Kitchen tools* Whisker, Slicing Knife Grater, Blender
Toys* Teddy bear, Rubiks cube Top, Plastic food
Furniture* Armchair, Sofa Dresser, Stool
Desks Office desk, Writing desk Reception desk, Cubicle desk
Houses Villa, Apartments Igloo, Canoe
Vehicles* Car, Bus Unicycle, Helicopter
Cars* Sedan, Hatchback Formula 1, Model T
Liquids Water, Milk Acetone, Soap
Drinks* Milk, Beer Cognac, Sake
Electronics* TV screen, Laptop Hair dryer, Shaver
Clothes* Shirt, Trousers Socks, Gloves
Games Puzzle, Chess Bowling, Super Mario
Music Musical note, The Beatles Mexican band, Accordion
Sports* Soccer, Basketball Bowling, Billiards
Religion Jesus, Western Wall Buddha, Praying man
Science* Test tubes, Atom Lecture, MRI
Conflicts IsraeliPalestinian Random couple argument
Symbols Peace symbol, David star Scouts symbol, Recycle symbol
Occupations* Judge, Policeman Fisherman, Violinist
Disasters 911 Plane crash, Tsunami Volcano eruption, Avalanche
Movies The Godfather, Cinema & Popcorn Cameraman, Script
Horror Wolf & full-moon, Hannibal Lecter Scared face, Creepy doll
Cartoons Mickey Mouse, The Simpsons Scooby-Doo, Hello Kitty
Events Wedding, Festival Graduation ceremony, Parade
Travel Passport & Suitcase, Backpackers Airport, Sunglasses
Health Heartbeat icon, Workout Nonsmoking, Granola
Hazard Slippery sign, Toxic (skull) sign Unstable bridge, Medusa
History Martin Luther King, Hiroshima Che Guevara, Mayan temples
Note. The 39 categories used in the student experiment; 20 categories for MTurks, indicated by *. Categories are placed in the first or second column
according to their being superordinate or basic level categories
Atten Percept Psychophys (2019) 81:28502872 2857
including a prototype (as nonmember or as member versus
nonmember same/different category). Of the 39 categories
used for Experiment 1, 20 were later tested in Experiment 2,
and only these were used in Experiment 3. For almost all the
20 categories, which were also tested in Experiment 2 (see
below), high typicality was confirmed; we discarded data for
the few discrepant images (<6% of trials). For the remaining
19 more conceptual categories, which were not tested in
Experiment 2 (and not used in Experiment 3), we depended
on examples from the literature (e.g., Iordan et al., 2016;
McCloskey & Glucksberg, 1978; Potter, Wyble, Pandav, &
Olejarczyk, 2010) and experimenter judgement for in-house
student participants (who came from the same cohort as ex-
perimenter NK). Note that if we err and choose nontypical
images as prototypes, this would add noise and reduce results
significance; thus, the results themselves confirm our choice.
For the entirely new MTurk tests, we used a different ap-
proach, depending on Experiment 2, as described below. We
purposely chose both basic and superordinate categories, as
well as conceptual categories, to broaden the potential impact
of our results.
Trial subtypes
Trial subtypes weredefined by the nature of the two testimage
objects vis-à-vis the sequence category (as in the low-level
tests; see the Introduction and Khayat & Hochstein, 2018).
Each SEEN test image could be of an object from the RSVP
sequence category (denoted SEENin) or the prototype of this
category (SEENprot). The NEW test image could be of an
object from the RSVP category (NEWin) or even its prototype
(NEWprot), but, in either case, not actually presented in the
sequence; alternatively, the NEW object image could be an
image of an object from a different category (NEWout).
Figure 3illustrates these image types. Each pair of test images
could be of one of five subtypes, listed in Table 3,(denoted
SEENprotNEWin, SEENinNEWin, SEENinNEWprot,
SEENinNEWout, or SEENprotNEWout). Each subtype
was tested for each category listed in Table 2.
Statistical tests and data analysis
Analysis of variance (ANOVA) tests with repeated measures
were conducted to verify that performance accuracy differ-
ences were due to the difficulty derived by effects emerging
from the different trial subtypes, rather than within-participant
differences in performance. For the two-way repeated-mea-
sures ANOVAs, testing student participant effects of SEEN
object typicality and NEW object category, we combined data
for NEW object same category, whether prototypical
(NEWprot) or not (NEWin); ttests (one-tailed) between the
averaged results of all participants for different subtype com-
binations were performed to investigate prototype and
boundary representations effects. Since it is difficult to re-
member all the sequence images, we expect participants to
correctly prefer as SEEN those test images with objects that
are prototypes of the sequence category (expected fraction
correct for SEENprotNEWin > for SEENinNEWin) and
mistakenly choose the NEW test image when it is the category
prototype, though not seen in the sequence (expected fraction
correct for SEENinNEWin > for SEENinNEWprot), and to
reject, as seen in the sequence, those that are of a different
category (fraction correct for SEENprotNEWout > for
SEENprotNEWin; and SEENinNEWout > SEENin
NEWin).
Results
The two basic measurements indicating observer performance
are accuracy rates and response time (RT) for each trial sub-
type, as shown for student participants in Fig. 4. The results by
trial subtype roughly resemble those from the low-level ex-
periment, demonstrated in Fig. 2b, with some effects even
more salient, as detailed below. Figure 5presents averaged
accuracy results across participants, sorted by subtype, isolat-
ing the three subtypes with both test image objects within the
sequence category (subtypes SEENprotNEWin, SEENin
NEWin, SEENinNEWprot), for student (Fig. 5a) and
MTurk participants (see Fig. 5b).
We performed a two-way repeated-measures ANOVA on
the Fig. 4results. The overall prototype effectthe effect of
one of the objects being the prototype of the category of the
objects presented in the sequencewas significant, F(1, 14) =
18.07, p< .001; the boundary effectthe effect of the non-
member being of another category than the sequence
objectswas highly significant, F(1, 14) = 298.64, p < .001,
and the interaction between them was significant, as well,
F(1,14) = 13.36, p< .005. The interaction effect suggests that
the prototype effect may be larger in some cases, as we shall
see in the following paragraph.
Prototype effect
The first factor to influence performance is the presence of
category prototypical objects (prototypes and most common
or familiar objects) in one of the test images. The presence of
typical exemplars influenced accuracy (% correct responses)
and RT, which together we call the prototype effect. As seen in
the three left bars of Fig. 4a and 5ab, prototype presence
affected accuracy: accuracy SEENprotNEWin > SEENin
NEWin > SEENinNEWprot. Prototype presence also affect-
ed response time (RT), as in Fig. 4b: RT correct choice of
member SEENprotNEWin < SEENinNEWin; RT incorrect
choice of nonmember SEENinNEWprot < SEENin
NEWin).
Atten Percept Psychophys (2019) 81:28502872
2858
It is possible that when including subtypes with
NEWout test images (i.e., images of an object of a different
category; subtypes SEENinNEWout and SEENprot
NEWout) in the above two-factor ANOVA calculation,
the effect of the presence of a different category
(NEWout) reduces the prototype effect. Thus, to test the
prototype effect alone, we conducted a one-way repeated-
measures ANOVA on the three subtypes, with test image
objects in the category boundaries (see Fig. 5). This one-
factor ANOVA showed a significant prototype effectstu-
dents: F(2, 28) = 11.78, p< .001; MTurk: F(2, 346) =
26.96, p< .001. We conclude that, as predicted, when
comparing trials containing only objects from the relevant
category (subtypes SEENprotNEWin, SEENinNEWin,
SEENinNEWprot), the prototype had a major influence
on observer response, which tended to attribute it as a
member of the RSVP sequence regardless of whether it
was or was not.
On the other hand, there is no significant difference be-
tween the case where the SEEN image object is prototypical
or not when the NEWobject is outside the category (accuracy
for SEENprotNEWout = 0.88 versus for SEENinNEWout =
0.86; p=.59;seeFig.4a). The boundary effect overrides the
prototype effect (leading to the interaction effect in the two-
way repeated-measures ANOVA, above).
We conclude that, due to limited attentional resources, par-
ticipants are unable to fully perceive and memorize all indi-
vidual objects, but still succeed in having a good representa-
tion of the category itself. This is striking, since the stimuli
were presented in RSVP manner, with brief periods between
stimuli. Nevertheless, observers were able to detect the se-
quence category and derive its prototype. They were success-
ful in both category and prototype determination for se-
quences that included basic level, subordinate, superordinate,
or even conceptual categories. They tend to relate the most
Table 3 Member recall test trial subtypes
SEEN test image (correct) NEW test image (incorrect) Expected performance
SEENprot NEWout Best
SEENin NEWout Better
SEENprot NEWin Better
SEENin NEWin Baseline
SEENin NEWprot Worse
Note. Each trial sequence of objects of a single category was followed by a pair of images of two objects, one a repeat of one of the object images in the
sequence,the SEEN image, and one an image of a NEW object. Choice of the SEEN image is correct, of the NEW image, incorrect. Test pairs of subtype
SEENinNEWin have both SEEN and NEW objects from the sequence category (in), but neither is the prototype. This is the baseline subtype against
which results from the other subtypes will be compared. In subtype SEENprotNEWin, the SEEN object is the category prototype, and the NEW object
is a category exemplar not shown in the sequence. If memory of the prototype is easier, we expect better performance for this subtype than for subtype
SEENinNEWin. In subtype SEENinNEWprot, theNEW image object is the category prototype, whichwas not shown in the sequence, and the SEEN
object is not the prototype. If there is false memory(i.e., after seeing a sequence of objects of a particular category, observers recallhaving seen the
category prototype), then they might choose, incorrectly, the unseen prototype rather than the seen object image. In subtypes SEENprotNEWout and
SEENinNEWout, the NEW object image is from another category (out), and the SEEN object is either the prototype (SEENprot) or is not the
prototype (SEENin). Here, we expect easy rejection of the NEW image object because it is of a different category. Irrespective of trial subtype,
participants sometime choose the SEEN test image because they remember seeing it in the sequence. Trial subtypes were presented in randomized
order without observers knowing about this classification
Fig. 4 High-level image memory performance by RSVP trial subtype
(Students). aAccuracy rates sorted by test image subtype (SEEN =
object image seen in trial sequence, NEW = object image not seen in
trial sequence). bResponse time measured for correct (choice of SEEN
image; green) and incorrect (choice of NEW image; red) responses, sorted
by test image subtype. (Color figure online)
Atten Percept Psychophys (2019) 81:28502872 2859
representative object (the prototype) to the category of the
presented object images and assume it was present in the se-
quence (see Fig. 5a: students; Fig. 5b: MTurks). We per-
formed post hoc ttests between the different subtypes to find
details of the effect, as shown in Fig. 5ab. The prototype
effect is clearly present when comparing the relevant trial
subtypes (SEENprotNEWin, SEENinNEWin, SEENin
NEWprot), which significantly differ from each other (stu-
dents: p< .05 for subtypes SEENinNEWin versus
SEENprotNEWin or SEENinNEWprot and p<.01for
SEENprotNEWin versus SEENinNEWprot; MTurks: p<
.001 for all comparisons). These subtypes create a staircase
shape from low performance of 0.54 ± 0.04 (MTurk: 0.64 ±
0.01; mean ± SE) proportion correct for SEENinNEWprot,
via 0.63 ± 0.02 (0.7 ± 0.008) correct for SEENinNEWin, to
best performance of 0.78 ± 0.02 (0.76 ± 0.01) correct for
SEENprotNEWin. We ask below if this is an all-or-none
prototype-or-not-prototype effect, or if it is a graded effect,
as objects are more or less typical of the category. Note that,
surprisingly, even when the prototype was not present in the
object sequence, it was often chosen as present when present-
ed as the NEW test image. Nevertheless, when choosing be-
tween a nonprototypical SEEN image and a prototypical
NEW image (SEENinNEWprot), having actually seen the
image in the sequence is slightly more important than typical-
ity (0.54 and 0.64 for students and MTurks, respectively; sig-
nificantly > .50). This is different than the results found for the
low-level feature set, as is easily seen in the proportion correct
for the SEENinNEWprot subtype (>.5) compared to the
analogous SEENinNEWmean subtype (<.5). We believe that
the difference derives from the greater observer memory for
images of real objects, compared to memory of absolute
values of simple features of abstract images (circle size, line
orientation, disc brightness).
We conclude that with a failure of coding all the individual
sequence images, due to brief image exposure times, the pres-
ence of prototype object images had a significant effect on the
responses, whether they were seen or new images of the
RSVP category. Along with these accuracy differences, an
analysis of the response times (RT) provides additional sup-
port for the conclusion that participants perceive prototypes as
ideal representatives of the category and rememberthese
whether they were present or not. In Fig. 6ab, RT is classified
into trials in which the NEW test image is correctly rejected
(Fig. 6ab, left green) or, incorrectly, chosen (Fig. 6ab, right
red) comparing when the NEW object either is or is not a
prototype. As expected, Fig. 6abshows that correct re-
sponses (green) are made faster than incorrect responses
(red), like the comparisons seen in Fig. 4b. The details show
further interesting comparisons, as follows. Analysis of the
correct RTs indicate that when participants did correctly chose
the nonprototype SEENin test image, they did so significantly
slower when the NEW image was a prototype (students:
1591 ms ± 125 ms; MTurk: 1364 ms ± 28 ms) than when
the NEW image was not a prototype (students: 1348 ms ±
46 ms; MTurk: 1319 ms ± 23ms; ttest p< .05), as displayed
in Fig. 6ab, left diamonds. In other words, not only were they
often manipulated to falsely pick the prototype as having been
seen in the sequence (see Fig. 5), even when they did manage
to choose a nonprototype SEEN image, their response was
delayed, as if the presence of the NEW image being a proto-
type (SEENinNEWprot) affected their confidence. In addi-
tion, choosing the correct SEEN object is faster when it is the
prototype (SEENprotNEWin versus SEENinNEWin and
Fig. 5 Category prototype object effect on accuracy. Proportion correct
for those subtypes for which both test objects are within the sequence
category: SEENinNEWprot, SEENinNEWin, and SEENprotNEWin;
ttests among the subtypes show significant differences, indicating the
expected prototype effect on observer judgment in membership tests, with
a preference to choose the object which matches the category prototype
(SEEN = object image seen in sequence, NEW = object image not seen in
sequence). aStudents. bMTurks. Significance indicated by *p<.05.**p
<.01.***p< .001
Atten Percept Psychophys (2019) 81:28502872
2860
SEENinNEWprot, see Fig. 4b, left three green bars).
Furthermore (Fig. 6ab, right red), choosing the NEW object,
incorrectly, is faster when it is the prototype than when it is not
(students: 1663 ms ± 150 ms versus 2015 ms ± 130 ms; ttest:
p=0.174,ns; MTurk: 1495 ms ± 41 ms versus 1557 ± 31 ms;
p<.05).
On the other hand, besides the prototype effect, there is
still some degree of recognition of test objects having been
seen in the sequence. Thus, as demonstrated in Fig. 6cd,
choosing the prototypical object is faster when it is a se-
quence member (correct: SEENprotNEWin; and
SEENprotNEWout for students; students: 1304 ms ± 50
ms; MTurk: 1288 ms ± 25 ms) than when it is not
(SEENinNEWprot incorrect; students: 1663 ms ± 150
ms;MTurk:1495ms±41ms;ttest: p<.05,p<.001,
respectively). Even choosing the nonprototypical seen im-
age is faster than choosing the typical new image (see Fig.
6ab, middle two diamonds; ttest: p= .061, p<.01).This
latter speed joins the greater accuracy (see above) to indi-
cate it is not a speedaccuracy trade-off.
Range/boundaries effect
The second statistic found for low-level sets is the range effect,
whose equivalent would be representation of category bound-
aries. A two-way repeated-measures ANOVA was performed
on accuracy and revealed a highly significant boundary effect,
as shown above, F(1, 14) = 298.64, p< .001. As with low-
level features, accuracy rates in trials of nonmember objects
outside of category boundaries (SEENprotNEWout and
SEENinNEWout; i.e., NEW objects from a different catego-
ry than the object sequence, were significantly higher, 0.87 ±
0.02, than in trials with both test objects within the category
range, SEENprotNEWin, SEENinNEWin, SEENin
NEWprot; 0.65 ± 0.02; p< .001), as seen in Fig. 7a.
This effect was observed also in response time measure-
ments for correct responses, as shown in Fig. 7b. Responses
were significantly faster for trials where the nonmember ob-
ject was outside category boundaries (i.e., belongs to a differ-
ent category, 1279 ms ± 54 ms), than in trials where both test
objects were from the category of the RSVP sequence
Fig. 6 Response Time Prototype Effect. aStudents: RT for each
combination for the NEW test element as prototype, not prototype,
correct, and incorrect trials. Green and red diamonds represent correct
and incorrect trials, respectively. Left: RT compared for correct trials
where the NEW test image object is the prototype of a category
(SEENinNEWprot) versus all other trials where it is not the prototype.
Right: RT compared for incorrect trials where the NEW test image is of
the prototype of a category (SEENinNEWprot) versus all other trials
where it is not of the prototype. Middle: RT compared for the NEW
object being the prototype and participants choosing this image,
incorrectly, or the nonprototype SEEN image, correctly. bSimilar graph
for MTurk participants. cStudents: RT comparison between trials with
participants picking prototype object images correctly (green bar) versus
incorrectly (red bar). dSimilar graph for MTurk participants. ***p=.001.
(Color figure online)
Atten Percept Psychophys (2019) 81:28502872 2861
(1476 ms ± 65 ms; p< .01). Taken together, the increase in
accuracy and decrease in RT indicate a consistent trend of
reducing task difficulty by introducing nonmember test ob-
jects from a different category, rather than a speedaccuracy
trade-off.
Experiment 2. Scoring object typicality
So far, we have compared results for category and set se-
quence member recall and effects of prototypemean and
boundariesrange edge on choice of member image in a 2-
AFC task. In addition, Khayat and Hochstein (2018) mea-
sured how these mean and range effects are graded with the
distance of the test item from the mean or from the range edge.
To complete and quantify the comparisons, we would like to
do the same for the prototype and category effects seen here.
To this end, we need a measure of the distance of our test
objects from their category prototype. (It would be nice to
measure how far away from a category are objects from dif-
ferent categories, but this seemed too difficult for the present
study.)
The current experiment was therefore designed to mea-
sure the subjective distance of objects from their category
prototype, and to learn for each category which object is
the prototype itself. To this end, we asked 50 MTurk par-
ticipants to choose one of two image objects as a member
of a previously named category, and used their response
speed as a measure of the closeness of the object to the
prototype. We will then use these results in Experiment 3
to measure the graded prototype effect. It has been well
documented that responses are faster for prototypes than
for non-prototypes (Ashby & Maddox, 1991,1994;
McCloskey & Glucksberg, 1979; Rips, Shoben, & Smith,
1973; Rosch, Simpson, & Miller, 1976). We note in the
Discussion that responses may also be faster for more fa-
miliar objects, and that there is debate concerning the rela-
tionship between familiarity and typicality.
Method
Stimuli and procedure
We present the name of a category in the middle of the screen
for 1s, (font: Arial 32, white), followed, after 1.0 s, by two test
images, one of an object belonging to the named category, and
one of a different category (attempting to choose objects that
were from a different category but not too far from the named
category; see Experiment 1, Method section). Images were
presented to the left and right of the center of the display, in
the middle half of the width and height of the screen. Images
remained present until observer response.
Observer task was to choose, by key press, the image with
an object that belongs to the named category. We hypothesize
that the closer the object is to the category prototype, the faster
will be the response, expecting participants to recognize pro-
totypical objects as members of the named category quicker
than they do atypical members. For example, participants will
recognize an apple as a fruit faster than a kiwi, a cow as a
mammal faster than a dolphin, and baseball as a sport faster
than mountain climbing.
We tested 50 Amazon Mechanical Turk participants
(MTurks). Participants performed two sessions of 300 trials/
session. They were tested on 20 categories, as indicated in
Tab le 2(starred categories), 10 categories per session, with
30 test objects for each category.
ab
Fig. 7 Range (category) effectwithin versus between category differ-
entiation (students). aAverage accuracy for subtypes SEENinNEWout
and SEENprotNEWout versus subtypes SEENprotNEWin, SEENin
NEWin, and SEENNEWprot. Observers were more accurate when the
nonmember test object was from a different category than that of the
RSVP sequence. bRT of correct trials was significantly faster when the
nonmember object belonged to a different category than when both test
objects belonged to the RSVP sequence category. **p=.01
Atten Percept Psychophys (2019) 81:28502872
2862
Results
As expected, response times varied among objects (maxi-
mum: 2.04 s; minimum: 0.65 s; mean range for 20 categories:
0.65 s), and there was significant correlation among partici-
pants (mean standard error between participants was 6%of the
RT).
Examples of categories and their objects are shown in
Fig. 8. For each category, four objects are shown, and for
each, the mean RT was measured for our 50 MTurk
observers.
We ranked the objects of each category (from 1 to 30) and
computed the mean RT for each rank over all 20 categories.
These average RTs were then normalized by: Normalized RT
=(RTminRT) / (maxRT minRT), where minRT and
maxRT are the minimum and maximum RTs for that category,
and (maxRT minRT) is the range of average (across partic-
ipant) response times for each category. Figure 9(blue sym-
bols) demonstrates the average normalized RT for each cate-
gory object rank. There is a high degree of across-category
similarity, evidenced by the small standard error among the
categories. Interestingly, RT dependence on rank is steeper at
the edges of the category objects, near the prototype (rank = 1)
and far from it (rank = 30). We also measured the across-
participant ranking and found small standard deviations (see
Fig. 9, red symbols). We shall now use this ranking as a typ-
icality index for each item in its category, to measure the
impact of typicality on object memory in the RSVP sequence
test.
Experiment 3. Graded typicality
Having derived a measure of the distance of each object from
its category prototypethe typicality indexwe now use this
index to measure the impact of typicality on memory of ob-
jects in a previously seen sequence. For low-level objects
(Khayat & Hochstein, 2018), it was easy to measure the dis-
tance of each element from the mean of the sequence since the
elements differed by a measurable feature (orientation,
brightness, size; see Fig. 1b). We found there, as shown in
Fig. 10b and d, that the mean effect is graded. That is, as the
member element is closer to the mean, so it is preferably
chosen as the member (see Fig. 10b). Similarly, as the non-
member is further from the mean, so it is rejected as not being
the member (see Fig. 10d). We now ask if this same rule
applies to category objects. We have seen the prototype effect
in Fig. 5as a preference to choose as the member objects that
are exactly the prototype of the category. Is this effect also
graded?
For Experiment 3, we tested MTurk participants (see
Experiment 1, Method section) with the 20 starred categories
in Table 2and tested in Experiment 2. We use the mean
across-participant RT found in Experiment 2 as the basis for
the typicality ranking of objects for Experiment 3. Note that
different MTurk participants were tested in Experiments 2 and
3 (Experiment 1 was with in-house student participants). For
Experiment 3, all objects presented in the test pairs were from
the same category as the previously presented sequence (only
bottom three subtypes of Table 3), so that we are now testing
the graded prototype effect, and not the range effect (seen in
Experiment 1; Figs. 4and 7).
Results
Figure 10 displays the graded prototype effect. We measure
the proportion correct,which is the probability of choosing the
member object as having been seen in the category sequence,
as a function of the typicality index of the member object (see
Fig. 10a). Typicality is ranked from 1 to 30, where 1 is the
closest to the prototype (i.e., the shortest average RT measured
in Experiment 2). Note the gradual decrease in choosing the
member as it is further from the prototype. Similarly, as the
nonmember is gradually further from typicalthat is, the
mean RT to this object was greater in Experiment 2, so this
object is more often rejected, and is less often chosen as the
member (see Fig. 10c).
Despite the Experiment 2 nonlinear dependence of typical-
ity rank on image RT, Fig. 10a and c data fit well a linear
regression. This may be because of the near linearity of the
Fig. 9curve, except at its extremes, and because Fig. 10a
averages over nonmember rank, Fig. 10covermemberrank,
and Fig. 11aoverboth.
The choice of an image is not dependent only on that im-
age, however, since there are always two images displayed
and we ask participants to choose between them. Thus, the
relative measure between the two images should determine
which image participants choose. Having found that sequence
member object closeness to the prototype and sequence non-
member distance from the category prototype both add to
correct choice of the member, we now plot choice accuracy
as a function of the difference between the distances of the
nonmember and the member. This is shown in Fig. 11a, where
we also show the parallel graph for low-level features
(Fig. 11b; from Khayat & Hochstein, 2018).
These graphs, including the high-level categorization
graphs, are not without noise. Noise comes from the random
second image in the membership tests, from interparticipant
differences, and from the very nature of our using RT as a
determinant for typicality. Nevertheless, the good fit to a sin-
gle trendline suggests that our conclusion is well founded, as
follows. When viewing a sequence of objects belonging to a
single category, observers often fail to recall the identity of
each object seen, and instead, when asked which of two ob-
jects was included in the sequence, depend, on recognition of
Atten Percept Psychophys (2019) 81:28502872 2863
the category seen, knowledge of the prototypical object, and
estimation of the distance of the two test objects from the
category prototype.
Discussion
The current results confirm and extend those of recent
studies suggesting that statistical representations general-
ize over a wide range of visual attributes, from simple
features to complex objects, giving accurate summaries
over space and time (Alvarez & Oliva, 2009;Ariely,
2001; Attarha & Moore, 2015; Chong & Treisman, 2003;
Gorea et al., 2014; Haberman & Whitney, 2009;Hubert-
Wallander & Boynton, 2015). This result is now extended
to object categories, as well. These efficient representa-
tions overcome severe capacity limitations of perceptual
resources (Alvarez & Oliva, 2008; Robitaille & Harris,
2011), and they are formed rapidly and early in conscious
visual representations (Chong & Treisman, 2003), without
focused attention (Alvarez & Oliva, 2008; Chong &
Treisman, 2005) and without conscious awareness of
Fig. 8 Examples of category objects and their associated response times.
Four example objects are shown for each of the categories of food, cars,
birds, animals, and clothing, with the mean RT over 47 observers. We
assume that shorter RTs are associated with objects that are closer to the
prototype, and use the RT ranking of objects for each category as a
measure for its typicality
Atten Percept Psychophys (2019) 81:28502872
2864
individual stimuli and their features (Demeyere,
Rzeskiewicz, Humphreys, & Humphreys, 2008;
Pavlovskaya, Soroker, Bonneh, & Hochstein, 2015).
Thus, their underlying computations play a fundamental
role in visual perception and the rapid extraction of infor-
mation from large and complex sources of data. In partic-
ular, we propose that categorization mimics set summary
statistics perception processes that share its characteristics.
Note that rapid gist perception does not imply low cortical
level representationon the contrary, it is the result of
rapid feed-forward computation along the visual hierarchy
(Hochstein & Ahissar, 2002).
Regarding high-level categories, we revealed two phenom-
ena that match those found for low-level features, by using a
similar experimental design for the two experiments: an
RSVP sequence followed by a 2-AFC experiment test of im-
age memory.
(1) Typicality effect: The typicality level of an object was
well represented, as it biased participantsdecision toward
choosing the more typical exemplar (of the presented catego-
ry) as the member of the RSVP sequence. The typicality effect
led to faster and more accurate responses for member test
items, and also to choice of the incorrect item, when it had
superior typicality (see Figs. 46and 1011). Thus, the more
typical object was chosen as present in the sequence, whether
it was or was not actually present there. The typicality effect is
similar to the set mean value effect found for low-level fea-
tures. (2) Boundary effect: Categorical boundary representa-
tion assisted participants in rejecting images with objects that
do not belong to the category of the RSVP sequence; they
therefore correctly chose the member image and achieved
higher performance levels in these trials (see Figs. 4and 7).
This effect is similar the set range edges effect.
Furthermore, using a dedicated response-time test to rank
the typicality of items within their category, we find that the
typicality effect is graded, similar to the set mean value effect
(see Figs. 10 and 11). The degree to which observers prefer-
entially choose category items as having been members of the
trial sequence is directly related to the degree of typicality of
the test items. Both member and nonmember items are chosen
more frequently as they are closer to prototypical; member
items, correctly, and nonmember items, incorrectly. In partic-
ular, the relative typicality of the member test item versus the
nonmember test item strongly affected observer choice of
which item they reported as member of the sequence (see
Fig. 11). Participants associated the more typical object to
the displayed RSVP sequence, regardless of whether the pro-
totype actually was or was not a member of the set. It is as if
when viewing the sequence of objects, they perceived the
category, but had only a poor representation of its individuals.
This is exactly what was found for set perception (Khayat &
Hochstein, 2018; Ward, Bear, & Scholl, 2016; but see Usher,
Bronfman, Talmor, Jacobson, & Eitam, 2018).
We propose that participants unconsciously considered
prototypes as better representatives of the categories than less
typical exemplars and correspondingly chose them as mem-
bers of the sequence, perhaps because prototypes usually con-
tain the most common attribute values shared among the cat-
egory members (Goldstone & Kersten, 2003;Rosch&
Mervis, 1975).
Fig. 9 Average normalized RT for each category object rank. Objects
were ranked from 1 to 30 for each category according to the RT in the
scoring object typicality test, where observers simply indicated which of
two objects belonged to a previously named category. We then
normalized the actual RTs, averaged over participants, and compare the
result with the ranking (blue). The fit of the two measures is very good,
with good agreement among participants. Also shown is the mean and
standard deviation across participants of the rank assigned to each objects
(red). The results match closely the mean RT data, and the across standard
deviation is small, confirming the methodology. (Color figure online)
Atten Percept Psychophys (2019) 81:28502872 2865
As in the low-level experiment, participants were not
informed about the categorical content of the RSVP se-
quences, and so they had no knowledge concerning the
involvement of prototypes, categories, and so forth, and
they only followed the instructions of an image memory
task. The similarity of the effects emerging from the two
experiments implies that statistical and categorical repre-
sentations are cognate phenomena that share perceptual
characteristics, and perhaps are generated by similar
computations.
a
b
c
d
Fig. 10 Graded prototype effect. aProportion correct as a function of the
typicality index of the member test object where typicality is ranked from
1 to 30 (1 closest to prototype). bSimilar graph for low-level feature
experiment (from Khayat & Hochstein, 2018). Proportion correct as func-
tion of member test-element distance from set mean. Note similar gradual
decrease in probability of choosing the member as it is further from the
prototype/mean. cProportion correct as a function of the typicality index
of the nonmember test object as it is gradually further from typical, so that
this object is more easily and more often rejected (i.e., less often chosen as
the sequence member). dSimilar graph for low-level feature experiment.
Note similarity between low-level feature and high-level categorization
effects
Atten Percept Psychophys (2019) 81:28502872
2866
Note that both the category prototype and boundary effects
are based on participantsimplicit categorization, extracted
from the images in the RSVP sequences.
The results indicate that they adjusted their responses to-
ward the relevant category, even though they were not guided
to take category information into consideration in the alleged
memory test. While participants concentrated on the RSVP
images themselves, it seems that category context extraction
overcame the cognitive abilities of memorizing the objects or
scenes presented by the images.
Nevertheless, we note that accuracy in this experiment was
superior to that in our previous set summary statistics experi-
ment (compare Figs. 4and 2; Khayat & Hochstein 2018). This
may well be due to accurate memory of some sequence items,
which is easier for object images than for abstract items (cir-
cles, disks, or line segments), which differ only in size, bright-
ness, or orientation. This result also confirms that participants
are trying to recall the actual objects displayed in the
sequencethey sometimes succeed in remembering them
and they are not consciously trying only to categorize the
images.
Categorical perception is often influenced by context
(Barsalou, 1987; Cheal & Rutherford, 2013; Joubert,
Rousselet, Fize, & Fabre-Thorpe, 2007;Koriat&Sorka,
2015,2017; Roth & Shoben, 1983). Water, for example,
may be associated with different categories, depending on
context. It is a drink, a liquid for bathing or cleaning, or the
medium of marine animals. Thus, the category to which par-
ticipants associated each sequence object would naturally be
affected by other sequence objects. We conclude that the cur-
rent categorization processes occurred rapidly and intuitively,
based on the variety of sequence objects, but also on earlier
processing of interactions between objects and their contexts
(Barsalou, 1987; Joubert et al., 2007; Koriat & Sorka, 2015,
2017; Roth & Shoben, 1983).
Differences between low-level parameter sets
and high-level categories
There are several differences between the low-level and the
high-level results that should be pointed out. For the low level,
we measured not only the graded mean effect, but also the
graded range effect (i.e., the gradual effect of the distance of
the presented nonmember element from the edge of the range
of the presented sequence). This range effect has its equivalent
in the boundary effect seen in Fig. 4. To extend this to a graded
effect would require measuring the distance between an object
of one category from the edgeof a different category. This is
beyond the scope of the current study.
y = 0.0287x + 0.5266
R² = 0.9575
0.3
0.4
0.5
0.6
0.7
0.8
0.9
-8 -7 -6 -5 -4 -3 -2 -1 0 1 2 3 4 5 6 7 8 9
Proporon correct
Difference between SEEN and NEW elements’ distance from mean
NEW closer to mean SEEN closer to mean
b
n=39
y = 0.004x + 0.7051
R² = 0.9434
0.5
0.6
0.7
0.8
0.9
-30 -25 -20 -15 -10 -5 0 5 10 15 20 25 30
Proporon correct
SEEN - NEW typicality rank difference
High-level objects difference - typicality effect
n=177
NEW more typical SEEN more typical
a
Low-level element difference - mean effect
Fig. 11 aAccuracy as a function of the difference between the distances
of the nonmember and the member objects from typicality (considering
that correct choice in the membership test depends on both the member
and the nonmember image distances from the prototypical. bParallel
graph for low-level features (Khayat & Hochstein, 2018)
Atten Percept Psychophys (2019) 81:28502872 2867
A second difference to be noted is that it is easier to remem-
ber particular pictures of objects than specific elements in a
sequence that differ only in a low-level feature (orientation,
size, or brightness). Thus, as mentioned above, performance
in the high-level test is superior overall. (Note performance
axis difference between Figs. 10a, c and b, d.)
Another significant difference between testing the low-
level set features and the high-level category objects is that
the set of low-level elements, and their range and mean, are
determined on the fly for each trial, by the sequence of stimuli
actually presented. In contrast, the high-level categoriesare, of
course, learned from life experience, and their prototype and
boundaries are known immediately when seeing the first ob-
ject in the sequence (or first few if the category is ambiguous).
Categorization is thus predetermined, and not a result of the
experience in the experiment itself. At the same time, there
may well be interparticipant differences in the way they cate-
gorize objects, and, in particular, in the specific objects that
they consider prototypical.
Related to the latter two differences is another. Categories
are often denoted and remembered by their name, introducing
a semantic element to the association of a variety of objects to
a single category. This is not so for the low-level features
studied previously. Nevertheless, recall that the world con-
tains, naturally and intrinsically, objects that cluster separately
in feature space, and thus categories that are language inde-
pendent (Goldstone & Hendrickson, 2010; Rosch, Mervis,
et al., 1976).
Implications for categorization processes
There is ongoing debate concerning category representation in
terms of the boundaries between neighboring categories, in
terms of a single prototype (category members resemble this
prototype more than they resemble other categoriesproto-
types), or in terms of a group of common exemplars (new
objects belong to the same category as the closest familiar
object). Our finding that participants respond on the basis of
both the mean and range of sets, and similarly on the basis of
the prototype and boundary of object categories may suggest a
hybrid categorization process model.
Concerning the single prototype versus multiple exemplar
theories, our results may support prototype theory, since we
find that participants choose test objects that are more proto-
typical, rather than recalling viewed exemplars. Nevertheless,
category prototypes may be a secondary readout of fuzzy rep-
resentations of multiple exemplars (see below).
We believe that the parallel found between set summary per-
ception and perception of categories suggests there might be a
common representation mechanism. We suggest that a popula-
tion code (Georgopoulos, Schwartz, & Kettner, 1986)might
underlie set representation of mean and range, and the same
may be true for category prototype and boundaries (Bonnasse-
Gahot & Nadal, 2008;Nicolelis,2001;Tajimaetal.,2016).
Observers clearly perceive not only the category of the
sequence objects but also their typicality (compare Evans &
Treisman, 2005;Potteretal.,2010). Furthermore, Benna and
Fusi (2019) suggested that related items (descendants from a
common ancestor in an ultrametric tree of items) may be effi-
ciently represented in a sparse and condensed manner by
representing their common ancestoror generator plus dif-
ferences of each item from it. Thus, representations of set and
category items might inherently include representation of
mean and prototype, respectively. Prototype theory is not
new, of course, but it is strengthened by the current finding
of the resemblance of categorization with set perception.
There is some debate concerning the relationship between ob-
ject familiarity and category typicality (Nosofsky, 1988;Palmeri&
Gauthier, 2004; Shen & Reingold, 2001). Responses are more
rapid for familiar objects (Wang, Cavanagh & Green, 1994; famil-
iar faces: Ramon, Caharelô, & Rossion, 2011; familiar
words: Glass, Cox & LeVine, 1974; familiar size: Konkle &
Oliva, 2012) or typical objects (Ashby & Maddox, 1991,1994;
McCloskey & Glucksberg, 1979;Ripsetal.,1973;Rosch,1973;
Rosch, Simpson, & Miller, 1976), but familiar objects are often
deemed more typical (Iordan, Green, Beck, & Fei-Fei, 2016;Malt
& Smith, 1982) and unfamiliar objects are quickly rejected from
category membership (Casey, 1992). Thus, our use of reaction
times for judging typicality may have included familiarity, and
our finding that participants chose more typical objects may
have included choice of more familiar objects. Nevertheless,
while Rosch (1973) found that categorization responses are faster
to prototypical objects, Ashby, Boynton, and Lee (1994) did not
find a meaningful correlation between response time and stimu-
lus familiaritywhen not related to category. In our experiments,
choosing the prototype or familiar object as having been SEEN
when it was not shown in the trial sequence, is surprising and not
expected based only on familiarity. Rather, such a result would be
consistent with a situation where sequence object representations
included a representation of their prototype (e.g., see Benna &
Fusi, 2019).
Our results resemble the DeeseRoedigerMcDermott
(DRM; Roediger & McDermott, 1995) finding that when pre-
sented with a list of related words, participants recall a
nonpresented lureword with the same frequency as the
presented words. In the DRM paradigm, participants study
lists of words (e.g., tired, bed, awake, rest, dream, night, blan-
ket, doze, slumber, snore, pillow, peace, yawn, drowsy) that
are related to a nonpresented lure word (e.g., sleep). On a later
test, participants often claim that they previously studied the
related lure words. Similarly, it was found that after learning a
set of distortions of a random dot pattern, participants learn the
undistorted patternthe prototypemore easily than a new
distortion, though only after a first viewing (Posner & Keele,
1968). These results may be added to the ensemble and
Atten Percept Psychophys (2019) 81:28502872
2868
categorization results, relating different situationssemantic
and perceptualwhere perceiving related items induces rep-
resentation and recall of the mean or prototype processes,
suggesting that similar processes may underlie them.
Such recall is referred to as falsememories, since false
recognition of the related lure words is indistinguishable from
true recognition of studied words (Gallo 2006; Schacter &
Addis, 2007). Our results, too, reflect falsememories, since
participants indicate recall of items that were not presented in
the sequence. This is equally true for our study of category
prototype recall and our studies, and those of many others, of
set ensemble presentation and recall of the set meaneven in
its absence from the presented sequence. Nevertheless, the
term false memoryis generally used in reference to recall
of events (Zaragoza, Hyman, & Chrobak, 2019) and narra-
tives (Frenda, Nichols, & Loftus, 2011) that did not occur or
were not narrated. Finding false memory of category proto-
type certainly extends this notion from abstract mean param-
eter (size, orientation, brightness, etc.) to more concrete ob-
jects and semantic categories, but this is still far from false
episodic memory. Further study is required to decide if these
different types of false memory are related, and if so, what is
the relationship between them.
Perceiving category exemplars in terms of the category
prototype may be the source of categorical priming (e.g.,
Fazio, Williams, & Powell, 2000;Ray,2008), whereby re-
sponses to unseen exemplars (and in particular to the category
prototype) are faster when primed by previously perceiving
another category exemplar. Interestingly, similar effects have
been found for sets (Marchant & de Fockert, 2009), and there
is even negative priming for unconscious viewing of single
unusual shapes (DeSchepper & Treisman, 1996).
Conclusions
We conclude that while observing the projected images, par-
ticipants first, implicitly generalized them into a category.
Then, at the membership test, they use this categorical context
to classify the probability of presence within the sequence of
the test images. That is, when visual memory capacity is in-
sufficient, then this implicit categorical context affects their
judgment. If indeed categorizations are executed by similar
computations as in statistical perception of the visual system,
then it is possible that these are only particular embodiments
of a general system, which efficiently determines our percep-
tion and behavior. It is especially poignant that set mean per-
ception and categorization, which help behaving in a too-rich
and too-complex environment by applying shortcuts to per-
ception, may share perceptual-computational mechanisms,
perhaps at different cortical levels. We have suggested that
the neural mechanism used is a population code
(Georgopoulos et al. 1986) that encodes both the mean and
the range of the stimulus set (Hochstein, 2016a,2016b;
Pavlovskaya, Soroker, Bonneh, & Hochstein, 2017a,2017b;
see also Brezis, Bronfman, Jacoby, Lavidor, & Usher, 2016;
Brezis, Bronfman, & Usher, 2018). Using a population code
to determine set mean answers the question of how the visual
system computes mean values without knowing values for
each element separately (whether represented when viewed
and forgotten, or never explicitly represented). Due to broad
tuning and overlap of neuron receptive fields, a population
code is necessarily used for perceiving individual element
values and may be used directly, with a broader range of
neurons over space and time, to perceive set mean values.
We now suggest the same type of population code may be
used for categorization. Category prototype and boundaries
could be the read out of fuzzy representations of multiple
exemplars. It has already been suggested that ensemble sum-
mary statistics might serve as the basis for rapid visual cate-
gorizations (Utochkin 2015).
A distinction was made between automatic, intuitive
global-attention scene gist perception, using vision at a glance,
versus explicit, focused-attention vision with scrutiny
(Hochstein & Ahissar, 2002). Gist is acquired automatically
and implicitly by bottom-up processing, and details are added
to explicit perception by further top-down guided processes.
The current study demonstrates that even when it is observers
intention to detect and remember the details of each image in a
sequencean intention that in this case often leads to fail-
urenevertheless, the automatic, implicit process of gist per-
ception succeeds in acquiring both set and category
information.
A question that still needs to be addressed is the cerebral
correlates of mechanisms underlying these processes. An in-
vestigation using physiological techniques (fMRI or EEG),
while participants perform behavioral tasks, as in the current
study, might indicate brain regions or electrophysiological
patterns of activity that are specific to systems that generate
these automatic representations. Such a study might also test
the notion that similar sites perform set mean and range per-
ception as well as categorization.
Acknowledgements We thank Yuri Maximov, the labs talented pro-
grammer, and lab comembers SafaAbassi and Miriam Carmeli.
Thanks to Stefano Fusi, Merav Ahissar, Udi Zohary, Israel Nelken,
Robert Shapley, Howard Hock, and Anne Treisman (of blessed memory),
for helpful discussions of earlier drafts of this paper. This study was
supported by a grant from the Israel Science Foundation (ISF).
The data for the experiments reported will be made available online,
and none of the experiments was preregistered.
Open Access This article is distributed under the terms of the Creative
Commons Attribution 4.0 International License (http://
creativecommons.org/licenses/by/4.0/), which permits unrestricted use,
distribution, and reproduction in any medium, provided you give appro-
priate credit to the original author(s) and the source, provide a link to the
Creative Commons license, and indicate if changes were made.
Atten Percept Psychophys (2019) 81:28502872 2869
References
Allik, J., Toom, M., Raidvee, A., Averin, K., & Kreegipuu, K. (2014).
Obligatory averaging in mean size perception. Vision Research,
101, 3440.
Alvarez, G. A., & Oliva, A. (2008). The representation of simple ensem-
ble visual features outside the focus of attention. Psychological
Science,19(4), 392398.
Alvarez, G. A., & Oliva, A. (2009). Spatial ensemble statistics are effi-
cient codes that can be represented with reduced attention.
Proceedings of the National Academy of Sciences of the United
States of America,106(18), 73457350.
Ariely, D. (2001). Seeing sets: Representation by statistical properties.
Psychological Science, 12(2), 157162.
Ashby, F. G., Boynton, G., & Lee, W. W. (1994). Categorization response
time with multidimensional stimuli. Perception & Psychophysics,
55, 1127.
Ashby, F. G., & Maddox, W. T. (1991). A response time theory of per-
ceptual independence. In J.-P. Doignon & J. C. Falmagne (Eds.),
Mathematical psychology: Current developments (pp. 389413).
New York, NY: Springer.
Ashby, F. G., & Maddox, W. T. (1994). A response time theory of sepa-
rability and integrality in speeded classification. Journal of
Mathematical Psychology, 38, 423466.
Ashby, F. G., & Maddox, W. T. (2011). Human category Learning 2.0.
Annals of the New York Academy of Sciences, 1224, 147161.
Ashby, F. G., & Valentin, V.V. (2017) Multiple systems of perceptual
category learning: Theory and cognitive tests. In H. Cohen & C.
Lefebvre (Eds.), Handbook of categorization in cognitive science
(2nd ed., pp 157188). Amsterdam, Netherlands: Elsevier.
Attarha, M., & Moore, C. M. (2015). The perceptual processing capacity
of summary statistics between and within feature dimensions.
Journal of Vision, 15(4), 9.
Barsalou, L. W. (1987). The instability of graded structure: Implications
for the nature of concepts. In U. Neisser (Ed.), Concepts and con-
ceptual development: Ecological and intellectual factors in catego-
rization (pp 101140). Cambridge, UK: Cambridge University
Press.
Bauer, B. (2009). Does Stevens's power law for brightness extend to
perceptual brightness averaging? The Psychological Record, 15(2),
171.
Bauer, B. (2015). A selective summary of visual averaging research and
issues up to 2000. Journal of Vision, 15(4):14, 115
Benna, M. K., & Fusi, S. (2019) Are place cells just memory cells?
Memory compression leads to spatial tuning and history depen-
dence. bioRxiv 624239. https://doi.org/10.1101/624239
Bonnasse-Gahot, L., & Nadal, J. P. (2008) Neural coding of categories:
Information efficiency and optimal population codes. Journal of
Computational Neuroscience, 25, 169187.
Brainard, D. (1997). The Psychophysics Toolbox. Spatial Vision, 10,
433436.
Brezis, N., Bronfman, Z. Z., Jacoby, N., Lavidor, M., & Usher, M. (2016).
Transcranial direct current stimulation over the parietal cortex im-
proves approximate numerical averaging. Journal of Cognitive
Neuroscience, 28(11), 114.
Brezis, N., Bronfman, Z. Z., & Usher, M. (2015). Adaptive spontaneous
transitions between two mechanisms of numerical averaging.
Scientific Reports, 5, 10415.
Brezis, N., Bronfman, Z., & Usher, M. (2018). A perceptual-like popu-
lation-coding mechanism of approximate numerical averaging.
Neural Computation, 30,428446.
Casey, P. J. (1992) A Reexamination of the roles of typicality and cate-
gory dominance in verifying category membership. Journal of
Experimental Psychology: Learning, Memory, and Cognition,
18(4), 823834.
Cheal, J. L., & Rutherford, M. D. (2013). Context-dependent categorical
perception of surprise. Perception, 42(3), 294301.
Chong, S. C., & Treisman, A. (2003). Representation of statistical prop-
erties. Vision Research, 43(4), 393404.
Chong, S. C., & Treisman, A. (2005). Attentional spread in the statistical
processing of visual displays. Attention, Perception, &
Psychophysics, 67,113.
Clapper, J. P. (2017) Alignability-based free categorization. Cognition,
162,87102.
Cohen, M. A., Dennett, D. C., & Kanwisher, N. (2016). What is the
bandwidth of perceptual experience? Trends in Cognitive
Sciences, 20(9), 324335.
Corbett, J. E., & Oriet, C. (2011). The whole is indeed more than the sum
of its parts: Perceptual averaging in the absence of individual item
representation. Acta Psychologica, 138(2), 289301.
Cowan, N. (2001). Metatheory of storage capacity limits. Behavioral and
Brain Sciences, 24(1), 154176.
Davis, T., & Love, B. C. (2010). Memory for category information is
idealized through contrast with competing options. Psychological
Science, 21,234242.
Demeyere, N., Rzeskiewicz, A., Humphreys, K. A., &Humphreys, G.W.
(2008). Automatic statistical processing of visual properties in
simultanagnosia. Neuropsychologia, 46(11), 28612864.
DeSchepper, B., & Treisman, A. (1996). Visual memory for novel
shapes: Implicit coding without attention. Learning, Memory, and
Cognition,22(1), 2747.
Evans, K. K., & Treisman, A. (2005). Perception of objects in natural
scenes: Is it really attention-free? Journal of Experimental
Psychology: Human Perception and Performance, 31,14761492.
Fabre-Thorpe, M. (2011). The characteristics and limits of rapid visual
categorization. Frontiers in Psychology, 2, 243, 112.
Fazio, R. H., Williams, C. J., & Powell, M. C. (2000) Measuring asso-
ciative strength: Category-item associations and their activation
from memory. Political Psychology, 21(1), 725.
Frenda, S. J., Nichols, R. M., & Loftus, E. F. (2011). Current issues and
advances in misinformation research. Current Directions in
Psychological Science, 20,2023.
Gallo, D. A. (2006). Associative illusions of memory. New York, NY:
Taylor & Francis.
Georgopoulos, A. P., Schwartz, A.B., & Kettner, R. E. (1986). Neuronal
population coding of movement direction. Science, 233, 1416
1419.
Glass, A. L., Cox, J., & LeVine, S. J. (1974). Distinguishing familiarity
from list search responses in a reaction time task. Bulletin of the
Psychonomic Society, 4, 105.
Goldstone, R. L., & Hendrickson, A. T. (2010). Categorical perception.
Wiley Interdisciplinary Reviews: Cognitive Science, 1(1), 6978.
Goldstone, R. L., & Kersten, A. (2003). Concepts and categorization. In I.
B. Weiner (Ed.), Handbook of psychology (pp. 597621). Hoboken,
NJ: Wiley.
Gorea, A., Belkoura, S., & Solomon, J. A. (2014). Summary statistics for
size over space and time. Journal of Vision, 14(9), 22, 114.
Haberman, J., & Whitney, D. (2007). Rapid extraction of mean emotion
and gender from sets of faces. Current Biology, 17(17), R751R753.
Haberman, J., & Whitney, D. (2009). Seeing the mean: ensemble coding
for sets of faces. Journal of Experimental Psychology: Human
Perception and Performance, 35(3), 718734.
Haberman,J., & Whitney, D. (2012). Ensembleperception: Summarizing
the scene and broadening the limits of visual processing. In J. Wolfe
& L. Robertson (Eds.), From perception to consciousness:
Searching with Anne Treisman (pp. 339349). New York, NY:
Oxford University Press.
Hammer, R., Diesendruck, G., Weinshall, D., & Hochstein, S. (2009).
The development of category learning strategies: What makes the
difference? Cognition, 112 (1), 105119.
Atten Percept Psychophys (2019) 81:28502872
2870
Hochstein, S. (2016b). How the brain represents statistical properties.
Perception, 45,272.
Hochstein, S. (2016a). The power of populations: How the brain repre-
sents features and summary statistics. Journal of Vision, 16(12),
1117.
Hochstein, S., & Ahissar, M. (2002). View from the top: Hierarchies and
reverse hierarchies in the visual system. Neuron, 36(5), 791804.
Hochstein, S., Khayat,N., Pavlovskaya, M., Bonneh, Y. S., & Soroker, N.
(2018). Set Summary perception, outlier pop out, and categoriza-
tion: A common underlying computation? Paper presented at the
41stEuropean Conference on Visual Perception, Trieste, Italy.
Hochstein, S., Khayat, N., Pavlovskaya, M., Bonneh, Y., Soroker, N., &
Fusi, S. (2019). Perceiving category set statistics on the fly. Journal
of Vision, 19.
Hochstein, S., Pavlovskaya, M., Bonneh, Y. S., & Soroker, N. (2015).
Global statistics are not neglected. Journal of Vision, 15(4), 7, 117.
Hochstein, S., Pavlovskaya, M., Bonneh, Y., & Soroker, N. (2018).
Comparing set summary statistics and outlier pop out in Vision.
Journal of Vision, 18(13/12), 113.
Hock, H. S., Gordon, G. P., & Whitehurst, R. (1974). Contextual rela-
tions: The influence of familiarity, physical plausibility, and belong-
ingness. Perception & Psychophysics, 16,48.
Hubert-Wallander, B., & Boynton, G. M. (2015). Not all summary statis-
tics are made equal: Evidence from extracting summaries across
time. Journal of Vision, 15(4), 5, 112.
Iordan, M. C., Greene, M. R., Beck, D. M., & Fei-Fei, L. (2015). Basic
level category structure emerges gradually across human ventral
visual cortex. Journal of Cognitive Neuroscience, 27(7), 129.
Iordan,M.C.,Greene,M.R.,Beck,D.M.,&Fei-Fei,L.(2016).
Typicality sharpens category representations in object-selective cor-
tex, Neuroimage, 134,170179.
Jackson-Nielsen, M., Cohen, M. A., & Pitts, M. A. (2017). Perception of
ensemble statistics requires attention. Consciousness and Cognition,
48,149160.
Joubert, O. R., Rousselet, G. A., Fize, D., & Fabre-Thorpe, M. (2007).
Processing scene context: Fast categorization and object interfer-
ence. Vision Research, 47(26), 32863297.
Khayat, N., & Hochstein, S. (2018). Perceiving set mean and range:
Automaticity and precision. Journal of Vision, 18(23), 114.
Konkle, T., & Oliva, A. (2012). A familiar-size Stroop effect: Real-world
size is an automatic property of object representation. Journal of
Experimental Psychology: Human Perception and Performance,
38(3), 561569.
Koriat, A., & Sorka, H. (2015). The construction of categorization judg-
ments: Using subjective confidence and response latency to test a
distributed model. Cognition, 134,2138.
Koriat, A., & Sorka, H. (2017). The construction of category membership
judgments: Towards a distributed model. In H. Cohen & C.
Lefebvre (Eds.), Handbook of categorization in cognitive science
(2nd ed., pp. 773794. Amsterdam, Netherlands: Elsevier.
Kriegeskorte, N., Mur, M., Ruff, D. A., Kiani, R., Bodurka, J., Esteky, H.,
Bandettini, P. A. (2008). Matching categorical object representa-
tions in inferior temporal cortex of man and monkey. Neuron, 60(6),
11261141.
Langlois, J. H., & Roggman, L. A. (1990). Attractive faces are only
average. Psychological Science, 1(2), 115121.
Luck, S. J., & Vogel, E. K. (1997). The capacity of visual working mem-
ory for features and conjunctions. Nature, 390(6657), 279281.
Maddox, W. T., & Ashby, F. G. (1993). Comparing decision bound and
exemplar models of categorization. Attention, Perception, &
Psychophysics, 53(1), 4970.
Malt, B. C., & Smith, E. E. (1982). The role of familiarity in determining
typicality. Memory & Cognition, 10,6975.
Marchant, A. P., & de Fockert, J. W. (2009). Priming by the mean repre-
sentation of a set, The Quarterly Journal of Experimental
Psychology, 62(10), 18891895.
McCloskey, M. E., & Glucksberg, S. (1978). Natural categories: Well
defined or fuzzy sets? Memory & Cognition, 6, 462472.
McCloskey, M., & Glucksberg, S. (1979). Decision processes in verify-
ing category membership statements: Implications for models of
semantic memory. Cognitive Psychology, 11 (1), 137.
Medin, D. L. (1989). Concepts and conceptual structure. American
Psychologist, 44(12), 14691481.
Medin, D. L., Altom, M. W., & Murphy, T. D. (1984). Given versus
induced category representations: Use of prototype and exemplar
information in classification. Journal of Experimental Psychology:
Learning, Memory, and Cognition, 10(3), 333352.
Morgan, M., Chubb, C., & Solomon, J. A. (2008). A dipperfunction for
texture discrimination based on orientation variance. Journal of
Vision, 8(11), 18.
Neumann, M. F.,Schweinberger, S. R., & Burton, A. M. (2013). Viewers
extract mean and individual identity from sets of famous faces.
Cognition, 128(1), 5663.
Nicolelis, M. A. L. (2001) Advances in neural population coding.
Progress in Brain Research, 130,3362.
Nomura, E. M., & Reber, P. J. (2008). A review of medial temporal lobe
and caudate contributions to visual category learning. Neuroscience
and Biobehavioral Reviews, 32(2), 279291.
Nosofsky, R. M. (1988). Exemplar-based accounts of relations between
classification, recognition, and typicality. Journal of Experimental
Psychology: Learning, Memory, and Cognition, 14(4), 700708.
Nosofsky, R. M. (2011). The generalized context model: An exemplar
model of classification. In E. M. Pothos, & A. J. Wills (Eds.), Formal
approaches in categorization (pp. 1839).NewYork,NY:
Cambridge University Press.
Nosofsky, R. M., Sanders, C. A., Gerdom, A., Douglas, B. J., &
McDaniel, M. A. (2017). Learning natural-science categories that
violate the family resemblance principle. PsychologicalScience, 28,
104114.
Oliva, A., & Torralba, A. (2006). Building the gist of a scene: The role of
global image features in recognition. In S. Martinez-Conde, S. L.
Macknik, L. M. Martinez, J.-M. Alonso, & P.U. Tse (Eds.), Progress
in brain research, visual perception, fundamentals of awareness:
Multi-sensory integration and high-order perception, 155B,2336.
Palmeri, T. J., & Gauthier, I. (2004). Visual object understanding. Nature
Neuroscience Reviews, 5,291304.
Pavlovskaya, M., Soroker, N., Bonneh, Y. S., & Hochstein, S. (2015).
Computing an average when part of the population is not perceived.
Journal of Cognitive Neuroscience, 27(7), 13971411.
Pavlovskaya, M., Soroker, N., Bonneh, Y., & Hochstein, S. (2017a).
Statistical averaging and deviant detection in heterogeneous arrays.
40th European Conference on Visual Perception Abstracts,40,160.
Pavlovskaya, M., Soroker, N., Bonneh, Y., & Hochstein, S. (2017b).
Statistical averaging and deviant detection may share mechanisms.
Washington, DC: Society for Neuroscience.
Posner, M. I., & Keele, S. W. (1968). On the genesis of abstract ideas.
Journal of Experimental Psychology, 77(3), 353363.
Posner, M. I., & Keele, S. W. (1970) Retention of abstract ideas. Journal
of Experimental Psychology, 83,304308.
Potter, M. C., & Hagmann, C. E. (2015). Banana or fruit? Detection and
recognition across categorical levels in RSVP. Psychonomic
Bulletin & Review, 22(2), 578585.
Potter, M. C., Wyble, B., Hagmann, C. E., & McCourt, E. S. (2014).
Detecting meaning in RSVP at 13ms per picture. Attention,
Perception, & Psychophysics, 76(2), 270279.
Potter, M. C., Wyble, B., Pandav, R., & Olejarczyk, J. (2010). Picture
detection in rapid serial visual presentation: Features or identity?
Journal of Experimental Psychology: Human Perception and
Performance, 36(6), 14861494.
Ray, S. (2008). An investigation of time course of category and semantic
priming. Journal of General Psychology, 135,2,133148.
Atten Percept Psychophys (2019) 81:28502872 2871
Ramon, M., Caharelô, S., & Rossion, B. (2011). The speed of recognition
of personally familiar faces. Perception, 40,437449.
Reed, S. K. (1972). Pattern recognition and categorization. Cognitive
Psychology, 3(3), 382407.
Rips, L. J., Shoben, E. J., & Smith, E. E. (1973). Semantic distance and
the verification of semantic relations. Journal of Verbal Learning and
Verbal Behavior, 12(1), 120.
Robitaille, N., & Harris, I. M. (2011). When more is less: Extraction of
summary statistics benefits from larger sets. Journal of Vision,
11(12), 18, 18.
Roediger, H. L., & McDermott, K. B. (1995). Creating false memories:
Remembering words not presented in lists. Journal of Experimental
Psychology: Learning, Memory, and Cognition, 21(4), 803814.
Rosch, E. (1973). Natural categories. Cognitive Psychology, 4(3), 328
350.
Rosch, E. (1999). Reclaiming cognition: The primacy of action, intention
and emotion. Journal of Consciousness Studies, 6(11/12), 6177.
Rosch, E. (2002). Principles of categorization. In D. Levitin (Ed.),
Foundations of cognitive psychology: Core readings (pp. 251
270). Cambridge, MA: MIT Press. (Original work published 1978)
Rosch, E., & Lloyd, B. B. (Eds.). (1978). Cognition and categorization.
Hillsdale, NJ: Erlbaum.
Rosch, E., & Mervis, C. B. (1975). Family resemblances: Studies in the
internal structure of categories. Cognitive Psychology, 7(4), 573
605.
Rosch, E., Mervis, C. B., Gray, W. D., Johnson, D. M., & Boyes-Braem,
P. (1976). Basic objects innatural categories. Cognitive Psychology,
8(3), 382439.
Rosch, E., Simpson, C., & Miller, R. S. (1976). Structural bases of typ-
icality effects. Journal of Experimental Psychology. Human
Perception and Performance, 2(4), 491502.
Roth, E. M., & Shoben, E. J. (1983). The effect of context on the structure
of categories. Cognitive Psychology, 15(3), 346378.
Schacter, D. L., & Addis, D. R. (2007). The cognitive neuroscience of
constructive memory: Remembering the past and imagining the fu-
ture. Philosophical Transactions of the Royal Society B, 362,773
786.
Shen, J., & Reingold, E. M. (2001) Visual search asymmetry: The influ-
ence of stimulus familiarity and low-level features. Perception &
Psychophysics, 63(3), 464475.
Sloutsky, V. M. (2003). The role of similarity in the development of
categorization. Trends in Cognitive Sciences, 7(6), 246251.
Smith, E. E., Langston, C., & Nisbett, R. (1992). The case for rules in
reasoning. Cognitive Science, 16,140.
Smith, J. D. (2014) Prototypes, exemplars, and the natural history of
categorization. Psychonomic Bulletin & Review, 21,312331.
Solomon, J. A. (2010). Visual discrimination of orientation statistics in
crowded anduncrowded arrays. Journal of Vision, 10(14), 19, 116.
Sweeny, T. D., Haroz, S., & Whitney, D. (2013). Perceiving group be-
havior: Sensitive ensemble coding mechanisms for biological mo-
tion of human crowds. Journal of Experimental Psychology: Human
Perception and Performance, 39(2), 329337.
Tajima, C. I., Tajima, S., Koida, K., Komatsu, H., Aihara, K., & Suzuki,
H. (2016). Population code dynamics in categorical perception.
Scientific Reports, 6, 22536. https://doi.org/10.1038/srep22536
Usher, M., Bronfman, Z. Z., Talmor, S., Jacobson, H., & Eitam, B.
(2018). Consciousness without report: Insights from summary sta-
tistics and inattention blindness. Philosophical Transactions of the
Royal Society B, 373, 20170354. https://doi.org/10.1098/rstb.2017.
035
Utochkin, I. S. (2015). Ensemble summary statistics as a basis for rapid
visual categorization. Journal of Vision, 15(4), 8, 114.
Wang, Q., Cavanagh, P., & Green, M. (1994) Familiarity and pop-out in
visual search. Perception & Psychophysics, 56(5), 495500.
Ward, E. J., Bear, A., & Scholl, B. J. (2016). Can you perceive ensembles
without perceiving individuals? The role of statistical perception in
determining whether awareness overflows access. Cognition, 152,
7886.
Yamanashi-Leib, A., Kosovicheva, A., & Whitney, D. (2016). Fast en-
semble representations for abstract visual impressions. Nature
Communications, 7, 13186, 110.
Zaragoza, M. S., Hyman, I., & Chrobak, Q. M. (2019). False memory. In
N. Brewer & A. B. Douglass (Eds.), Psychological science and the
law (pp 182207). New York, NY: Guilford Press.
PublishersnoteSpringer Nature remains neutral with regard to
jurisdictional claims in published maps and institutional affiliations.
Atten Percept Psychophys (2019) 81:28502872
2872
... Additionally, (one-tailed) t-tests between the averaged results over participants for different subtype combinations were performed to investigate ancestor and range representation effects. Some of these experiments have been reported previously in meeting abstract or brief communication format (e.g., Hochstein, 2019;Hochstein et al., 2018Hochstein et al., , 2020Hochstein, Khayat, Pavlovskaya, Bonneh, Soroker, & Fusi, 2019;Khayat & Hochstein, 2019b). ...
Article
Full-text available
Perception, representation, and memory of ensemble statistics has attracted growing interest. Studies found that, at different abstraction levels, the brain represents similar items as unified percepts. We found that global ensemble perception is automatic and unconscious, affecting later perceptual judgments regarding individual member items. Implicit effects of set mean and range for low-level feature ensembles (size, orientation, brightness) were replicated for high-level category objects. This similarity suggests that analogous mechanisms underlie these extreme levels of abstraction. Here, we bridge the span between visual features and semantic object categories using the identical implicit perception experimental paradigm for intermediate novel visual-shape categories, constructing ensemble exemplars by introducing systematic variations of a central category base or ancestor. In five experiments, with different item variability, we test automatic representation of ensemble category characteristics and its effect on a subsequent memory task. Results show that observer representation of ensembles includes the group’s central shape, category ancestor (progenitor), or group mean. Observers also easily reject memory of shapes belonging to different categories, i.e. originating from different ancestors. We conclude that complex categories, like simple visual form ensembles, are represented in terms of statistics including a central object, as well as category boundaries. We refer to the model proposed by Benna and Fusi ( bioRxiv 624239, 2019) that memory representation is compressed when related elements are represented by identifying their ancestor and each one’s difference from it. We suggest that ensemble mean perception, like category prototype extraction, might reflect employment at different representation levels of an essential, general representation mechanism.
... Ensemble perception has been studied for basic parameters, including size (Allik, Toom, Raidvee, Averin, & Kreegipuu, 2014;Ariely, 2001;Corbett & Oriet, 2011;Morgan, Chubb, & Solomon, 2008;Solomon, 2010), orientation (Alvarez & Oliva, 2009;Hochstein, Pavlovskaya, Bonneh, & Soroker, 2018), brightness (Bauer, 2009), spatial position (Alvarez & Oliva, 2008), and speed and direction of motion (Sweeny, Haroz, & Whitney, 2013). Summary statistics perception appears to be a general mechanism operating on various stimulus attributes, including these low-level parameters as noted above, and complex characteristics, such as facial expression or emotion and gender (Haberman & Whitney, 2007;Haberman & Whitney, 2009;Neumann, Schweinberger, & Burton, 2013), object lifelikeness (Yamanashi Leib, Kosovicheva, & Whitney, 2016), biological motion of human crowds (Sweeny, Haroz, & Whitney, 2013), numerical averaging (Brezis, Bronfman, & Usher, 2015) and even category membership Hochstein, Khayat, Pavlovskaya, Bonneh, Soroker, & Fusi, 2019) for recent reviews, see Bauer, 2015;Cohen et al., 2016;Haberman & Whitney, 2012;Hochstein, Pavlovskaya, Bonneh, & Soroker, 2015;Whitney & Yamanashi Leib, 2018; and an upcoming Attention, Perception, and Psychophysics special issue). ...
Article
Full-text available
Previous studies have demonstrated a complex relationship between ensemble perception and outlier detection. We presented two array of heterogeneously oriented stimulus bars and different mean orientations and/or a bar with an outlier orientation, asking participants to discriminate the mean orientations or detect the outlier. Perceptual learning was found in every case, with improved performance accuracy and speeded responses. Testing for improved accuracy through cross-task transfer, we found considerable transfer from training outlier detection to mean discrimination performance, and none in the opposite direction. Implicit learning in terms of increased accuracy was not found in either direction when participants performed one task, and the second task's stimulus features were present. Reaction time improvement was found to transfer in all cases. This study adds to the already broad knowledge concerning perceptual learning and cross-task transfer of training effects.
... The situation with categorization is unlike that with sets, where we derive the set mean, on the fly, as we are presented with set members. Instead, when encountering an object (or group of objects belonging to a single category), we know the category to which it belongs, and we also know what is the prototype of that category and the category boundaries; there is no need, and no possibility, of deriving anew the category, prototype, and boundaries of a group of familiar objects (though we can learn new categories of unfamiliar objects; see Hochstein et al., 2019). Furthermore, categories may be learned and recognized semantically, while the basic features of sets are often nonsemantic. ...
Article
Full-text available
Two cognitive processes have been explored that compensate for the limited information that can be perceived and remembered at any given moment. The first parsimonious cognitive process is object categorization. We naturally relate objects to their category, assume they share relevant category properties, often disregarding irrelevant characteristics. Another scene organizing mechanism is representing aspects of the visual world in terms of summary statistics. Spreading attention over a group of objects with some similarity, one perceives an ensemble representation of the group. Without encoding detailed information of individuals, observers process summary data concerning the group, including set mean for various features (from circle size to face expression). Just as categorization may include/depend on prototype and intercategory boundaries, so set perception includes property mean and range. We now explore common features of these processes. We previously investigated summary perception of low-level features with a rapid serial visual presentation (RSVP) paradigm and found that participants perceive both the mean and range extremes of stimulus sets, automatically, implicitly, and on-the-fly, for each RSVP sequence, independently. We now use the same experimental paradigm to test category representation of high-level objects. We find participants perceive categorical characteristics better than they code individual elements. We relate category prototype to set mean and same/different category to in/out-of-range elements, defining a direct parallel between low-level set perception and high-level categorization. The implicit effects of mean or prototype and set or category boundaries are very similar. We suggest that object categorization may share perceptual-computational mechanisms with set summary statistics perception.
Article
Full-text available
Visual scenes are too complex to perceive immediately in all their details. Two strategies (among others) have been suggested as providing shortcuts for evaluating scene gist before its details: (a) Scene summary statistics provide average values that often suffice for judging sets of objects and acting in their environment. Set summary perception spans simple/complex dimensions (circle size, face emotion), various statistics (mean, variance, range), and separate statistics for discernible sets. (b) Related to set summary perception is detecting outliers from sets, called "pop out," which allows rapid perception of presence and properties of unusual, and thus, possibly salient items in the scene. To understand better visual system mechanisms underlying these two set-related perceptual phenomena, we now study their properties and the relationship between them. We present observers with two clouds of bars with distributed orientations and ask them to discriminate their mean orientations, reporting which cloud is oriented more clockwise, on average. In the second experiment, the two clouds had the same mean orientation, but one had a bar with an outlier orientation, which observers detected. We find that cloud mean orientation discrimination depends mainly on the difference in means, whereas outlier detection depends mainly on the distance of the outlier orientation from the edge of the cloud orientation range. Neither percept depends largely on the range of the set itself. A unified model of a population-code mechanism underlying these two phenomena is discussed.
Article
Full-text available
The classification of objects to natural categories displays a great deal of cross-person consensus and within-person consistency. At the same, categorization also exhibits some degree of within-person instability and cross-person variability. We attempted to gain insight into the stable and variable contributions to category membership judgment by examining confidence judgments and response latency for one's decision whether an object belongs or does not belong to a given category. According to an extension of the Self-Consistency Model (SCM) (Koriat, 2012), category membership decisions are constructed on the fly on the basis of a small set of cues sampled sequentially from a population of cues associated with the object-category item. This population is largely shared by participants with the same background. The decision is based on the balance of evidence in favor of a positive or a negative response, and confidence is based on the consistency with which that decision was supported across the accessed cues. The results confirmed several predictions: (a) Consensual responses were endorsed with higher confidence and shorter response latency than nonconsensual responses with the differences between the two types of responses increasing with item consensus -- the proportion of participants who made the majority response for the item. (b) When the task was repeated several times, confidence and response speed were higher for each participant’s more frequent decision than for the less frequent decision. (c) Results suggested that confidence in a category membership decision reflects the participant’s assessment of the likelihood that the same decision will be reached in future encounters with the item. (d) Finally, the context that was provided for the category membership decision was found to bias the decision reached, but confidence also changed correspondingly, suggesting that context affected the sampling of cues underlying the decision. Altogether, the results provide support for the model, and indicate that confidence and response latency can track the sources of stability and variability in category membership decisions.
Article
Full-text available
Much evidence suggests that real-world natural kinds are based on overall similarity or family resemblance, but people often appear surprisingly insensitive to family resemblance in laboratory studies of sorting or free categorization. In such experiments, all stimuli generally vary along the same discretely-varying dimensions and family resemblance is defined in terms of the proportion of matching or mismatching values along those dimensions. This article argues for an alternative conception of family resemblance based on structural alignability, i.e., whether objects have corresponding parts-in-relations that can provide the basis for a shared schema or conceptual model. Five experiments using two new free categorization tasks demonstrate that structural alignment, even without specific matching parts, is sufficient for people to perceive objects as essentially similar and group them into common family-level categories. Importantly, the experiments demonstrate that this categorization is based on abstract alignment rather than shared parts or features, because when the parts of the individual objects are randomly rearranged, eliminating their shared spatial structure, people no longer perceive them as belonging to a common category. These results suggest that people do construct perceptual categories on the basis of overall similarity, at least when similarity is defined in terms of spatial correspondence or alignability rather than individual shared parts or features.
Article
Full-text available
Much of the richness of perception is conveyed by implicit, rather than image or feature-level, information. The perception of animacy or lifelikeness of objects, for example, cannot be predicted from image level properties alone. Instead, perceiving lifelikeness seems to be an inferential process and one might expect it to be cognitively demanding and serial rather than fast and automatic. If perceptual mechanisms exist to represent lifelikeness, then observers should be able to perceive this information quickly and reliably, and should be able to perceive the lifelikeness of crowds of objects. Here, we report that observers are highly sensitive to the lifelikeness of random objects and even groups of objects. Observers' percepts of crowd lifelikeness are well predicted by independent observers' lifelikeness judgements of the individual objects comprising that crowd. We demonstrate that visual impressions of abstract dimensions can be achieved with summary statistical representations, which underlie our rich perceptual experience.
Preprint
The observation of place cells has suggested that the hippocampus plays a special role in encoding spatial information. However, place cell responses are modulated by several non-spatial variables, and reported to be rather unstable. Here we propose a memory model of the hippocampus that provides a novel interpretation of place cells consistent with these observations. We hypothesize that the hippocampus is a memory device that takes advantage of the correlations between sensory experiences to generate compressed representations of the episodes that are stored in memory. A simple neural network model that can efficiently compress information naturally produces place cells that are similar to those observed in experiments. It predicts that the activity of these cells is variable and that the fluctuations of the place fields encode information about the recent history of sensory experiences. Place cells may simply be a consequence of a memory compression process implemented in the hippocampus. Significance Statement Numerous studies on humans revealed the importance of the hippocampus in memory formation. The rodent literature instead focused on the spatial representations that are observed in navigation experiments. Here we propose a simple model of the hippocampus that reconciles the main findings of the human and rodent studies. The model assumes that the hippocampus is a memory system that generates compressed representations of sensory experiences using previously acquired knowledge about the statistics of the world. These experiences can then be memorized more efficiently. The sensory experiences during the exploration of an environment, when compressed by the hippocampus, lead naturally to spatial representations similar to those observed in rodent studies and to the emergence of place cells.
Article
To compensate for the limited visual information that can be perceived and remembered at any given moment, many aspects of the visual world are represented as summary statistics. We acquire ensemble representations of element groups as a whole, spreading attention over objects, for which we encode no detailed information. Previous studies found that different features of items (from size/orientation to facial expression/biological motion) are summarized to their mean, over space or time. Summarizing is economical, saving time and energy when the environment is too rich and complex to encode each stimulus separately. We investigated set perception using rapid serial visual presentation sequences. Following each sequence, participants viewed two stimuli, member and nonmember, indicating the member. Sometimes, unbeknownst to participants, one stimulus was the set mean, and or the nonmember was outside the set range. Participants preferentially chose stimuli at/near the mean, a "mean effect," and more easily rejected out-of-range stimuli, a "range effect." Performance improved with member proximity to the mean and nonmember distance from set mean and edge, though they were instructed only to remember presented stimuli. We conclude that participants automatically encode both mean and range boundaries of stimulus sets, avoiding capacity limits and speeding perceptual decisions.