Blinking Bird Brains: A Timing Specific Deficit in Auditory Learning in Quail Hatchlings

Abstract

Contingency perception is critical during infancy, providing the basis for the continuous learning (and re-learning) of the relation between the developing body, self-produced actions, and the environment. Nevertheless, relatively few studies have systematically examined the spatiotemporal parameters that optimize learning during development. Here, we present a series of experiments exploring a novel timing-specific deficit in auditory learning in an animal model, the bobwhite quail (Colinus virginianus). In this paradigm, chicks vocalize to hear playback of an unfamiliar maternal call and are later tested for their filial preference for that call over a novel maternal call. Rather than a simple, hyperbolic decline in learning with increased delay between chick vocalization and playback, we found a window—450-900 ms after chicks ceased vocalizing—in which chicks appeared to have difficulty learning and forming a preference for the maternal call. This deficit nonetheless occurred only when the spatial location of call playback switched semi-randomly during training, suggesting an attentional explanation for this deficit. Our findings indicate that optimization of the temporal parameters in operant paradigms with infants can be complex, particularly if tasks requiring the switching of attention between spatial locations. Our findings may thus be instructive for other developmental research with infants employing operant components.

Full text

Blinking Bird Brains: A Timing Specific Deficit
in Auditory Learning in Quail Hatchlings
Christopher Harshaw
Department of Psychological & Brain Sciences
Indiana University
Robert Lickliter
Department of Psychology
Florida International University 1;2
Contingency perception is critical during infancy, providing the basis for the continuous
learning (and re-learning) of the relation between the developing body, self-produced
actions, and the environment. Nevertheless, relatively few studies have systematically
examined the spatiotemporal parameters that optimize learning during development. Here,
we present a series of experiments exploring a novel timing-specific deficit in auditory
learning in an animal model, the bobwhite quail (Colinus virginianus). In this paradigm,
chicks vocalize to hear playback of an unfamiliar maternal call and are later tested for
their filial preference for that call over a novel maternal call. Rather than a simple, hyper-
bolic decline in learning with increased delay between chick vocalization and playback, we
found a window—450–900 msec after chicks ceased vocalizing—in which chicks appeared
to have difficulty learning and forming a preference for the maternal call. This deficit
nonetheless occurred only when the spatial location of call playback switched semi-ran-
domly during training, suggesting an attentional explanation for this deficit. Our findings
indicate that optimization of the temporal parameters in operant paradigms with infants
can be complex, particularly if tasks requiring the switching of attention between spatial
locations. Our findings may thus be instructive for other developmental research with
infants employing operant components.
Infants are exquisitely sensitive to contingencies between their actions and sensory
feedback (e.g., Bahrick & Watson, 1985; Bloom, 1977; Filippetti, Johnson, Lloyd-Fox,
Dragovic, & Farroni, 2013; Goldstein, Schwade, & Bornstein, 2009; Rochat & Mor-
gan, 1995). Infants often responding to such contingencies—particularly “imperfect”
ones, wherein interoceptive and exteroceptive feedback diverge—with attentiveness,
enthusiasm, and displays of positive affect (see Watson, 1972). Contingencies provide
Correspondence should be sent to Christopher Harshaw, Department of Psychological & Brain Sciences,
Indiana University, 1101 E. 10th Street, Bloomington, IN, 47405, USA. E-mail: chrisharshaw@yahoo.com
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
I N F A 12139
Dispatch: 17.2.16 CE: Liyagath Ali
Date of receipt of final manuscript: 17 July
2015
Revised 4 February 2016
Accepted 5 February 2016
Journal Code Manuscript No.
No. of pages: 28 PE: Pavithra
Infancy, 1–28, 2016
Copyright ©International Congress of Infant Studies (ICIS)
ISSN: 1525-0008 print /1532-7078 online
DOI: 10.1111/infa.12139

infants not just a source of pleasure but, by their very nature, a critical source of infor-
mation about the statistics of their sensory worlds (Goldstein, Schwade, Briesch, &
Syal, 2010; Goldstein, Waterfall, et al., 2010). For example, infants can use contingen-
cies to distinguish social from non-social stimulation (Watson, 1972, 1979) and such
information may serve as scaffolding for an emerging sense of “self” during infancy
(see Bahrick, 1995, 2013; Gergely & Watson, 1999). The ability to discriminate
between social and non-social objects and to closely attend to social (and potentially
social) contingencies is moreover critical for infants, given that caregivers and other
social partners are often the major source of both needed resources—like food and
shelter—as well as crucial information about the world. For example, contingencies
play a critical role in language development during infancy and childhood (see Gold-
stein & Schwade, 2010). Caregivers not only respond contingently to the pre-linguistic
babbling of infants (Bornstein, Putnick, Cote, Haynes, & Suwalsky, 2015; Gros-Louis,
West, Goldstein, & King, 2006), but infants are able to utilize such feedback to pro-
duce more sophisticated, language-like utterances, even when the feedback provide by
caregivers is non-vocal (Goldstein, King, & West, 2003; Goldstein & Schwade, 2008).
Similar contingencies appear to continue to scaffold language learning throughout
childhood (e.g., Roseberry, Hirsh-Pasek, & Golinkoff, 2014).
For most avian species, infancy is similarly characterized by both high levels of par-
ental involvement and an abundance of social contingencies. Like humans, young birds
appear to utilize such contingencies to actively acquire knowledge of their worlds (e.g.,
Evans, 1991; Schneider & Harshaw, 2007; Schneider & Lickliter, 2010; West & King,
1988). For example, in precocial species such as ducks and quail, chicks begin producing
a variety of sounds, including chirps and ‘peeps’ while still in the egg (i.e., as embryos;
see Rumpf & Tzschentke, 2010). These embryonic vocalizations not only shape these
birds’ later ability to recognize the calls of their own species (Gottlieb, 1971, 1997), but
adults begin to respond to such vocalizations, beginning even prior to hatch (e.g., Got-
tlieb, 1963; Impekoven, 1973; Tuculescu & Griswold, 1983). This pattern of reciprocal
exchange, at least superficially resembling “conversation,” continues and develops
throughout the early postnatal period (e.g., Collias, 1952; Rumpf & Tzschentke, 2010;
Stokes, 1967). Despite such facts, studies investigating the process by which young pre-
cocial birds acquire preferences for social stimuli (i.e., filial imprinting)—exemplified by
the work of Konrad Lorenz (Lorenz, 1937)—historically focused almost exclusively on
visual characteristics of the mother or imprinting stimulus (cf. Gottlieb, 1963) while
employing almost exclusively passive exposure paradigms, devoid of ecologically and
developmentally relevant contingencies (Harshaw & Lickliter, 2007; Lickliter & Har-
shaw, 2010). The auditory version of this approach generally involved exposing embryos
and neonates to several hours of stimulation prenatally to induce a preference for a par-
ticular maternal call postnatally (e.g., Gottlieb, 1987; Lickliter & Hellewell, 1992).
In a recent series of studies, we explored a more ecologically grounded, interactive
approach to the emergence of filial preferences in hatchlings of a precocial bird, the
Northern bobwhite quail (Colinus virginianus). When precocial chicks such as the bob-
white are separated from their brood, they typically vocalize vigorously—giving off a
rapid string of “peeps”—which function to maintain contact with parents and brood-
mates when out of visual contact (Stoumbos, 1990). We took advantage of this propen-
sity and gave day-old chicks the opportunity to “peep” to trigger contingent (CON)
playback of a bobwhite maternal call while isolated from their brood (Harshaw & Lick-
liter, 2007, 2011; Harshaw, Tourgeman, & Lickliter, 2008). These chicks received a sin-
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
2HARSHAW &LICKLITER

gle playback of a previously unfamiliar bobwhite maternal call each time they ceased
vocalizing, during a single 5 min training session at 24 h of age, and were tested for their
preference between this familiar and an unfamiliar maternal call 24 h later (Harshaw &
Lickliter, 2007). Not only did chicks acquire preferences for a previously novel call more
rapidly with this approach—in only 5 min compared to 240–480 min in passive-expo-
sure paradigms (Foush
ee & Lickliter, 2002; Lickliter & Hellewell, 1992)—but CON
exposure to a foreign, species-atypical call eliminated the usual strong preference for a
species-typical, bobwhite maternal call over the call of other species (Harshaw & Lick-
liter, 2011; Harshaw et al., 2008). Adding a small amount of prenatal exposure to the
species-atypical call prior to postnatal contingent training with the same call moreover
induced a reversal of chicks’ usual species-typical preference, with such chicks coming to
significantly prefer the foreign call to that of their own species (Harshaw & Lickliter,
2011). Yoked, non-contingent (YNC) exposure had no such effects in these studies (Har-
shaw & Lickliter, 2007, 2011; Harshaw et al., 2008). Our results thus demonstrate that
social contingencies can play a powerful role in driving the canalization of species-typi-
cal social preferences during early development (Lickliter & Harshaw, 2010).
Despite the overall superiority of contingent (CON) compared to yoked, non-con-
tingent (YNC) exposure for inducing preferences in young birds, some chicks receiving
YNC exposure in our experiments nevertheless individually displayed preferences for
the exposed maternal call as strong as those of CON chicks (Harshaw & Lickliter,
2007; Harshaw et al., 2008). One possible explanation for this effect was that, despite
the lack of programmed or systematic contingency, some chicks receiving YNC expo-
sure were “superstitiously” associating their frequent contact peeps with the yoked
playbacks, due to incidental temporal proximity (cf. (Skinner, 1948; Timberlake &
Lucas, 1985). A rigorous assessment of this possibility required us to first determine
the temporal envelope within which chicks respond to an auditory stimulus as if it
occurs because of their own behavior—that is, ‘When does a young bird perceive a
contingency?’ We thus began what we expected would be a simple parametric explo-
ration of day-old chicks’ auditory learning in response to delayed playback of a mater-
nal call, using a procedure otherwise identical to that used in our previous studies.
The negatively accelerated hyperbolic relation between delay and learning or reward
value of contingently presented stimuli has been well characterized (see Mazur, 2001)
and widely generalized (e.g., Blackburn & El-Deredy, 2013; Kirby & Marakovi
c, 1996;
Odum, Baumann, & Rimington, 2006). We thus expected to find a straightforward,
hyperbolic decline in learning (Mazur, 2001; Takahashi, Ikeda, & Hasegawa, 2007)
with increased delay between chicks’ “peeps” and playback of a maternal call. As we
will describe, chicks instead showed a previously undocumented timing-specific deficit
or “blink” in learning with delays specifically in the range of 450–900 msec between
their contact vocalizations and maternal call playback: chicks displayed good learning
of the maternal call on either side of this window and a gradual decline after ~1,500–
2,000 msec. A quirk of our experimental design—the semi-random switching of play-
back between two locations from playback to playback, which we employed to prevent
the emergence of spatial bias—was eventually found to be the apparent trigger of the
deficit. The study presented here describes our initial experiment and documentation of
this deficit or “blink” in learning in quail chicks as well as several follow-up experi-
ments aimed at both exploring the generality of and elucidating the mechanisms
underlying this deficit. Although more experiments are needed, we argue that an atten-
tional account, drawing on comparisons with human work, provides the most parsimo-
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
TIMING-SPECIFIC DEFICIT IN AUDITORY LEARNING 3

nious explanation of our findings. Given the utility of operant approaches (e.g., in con-
jugate reinforcement) and paradigms with embedded operants in developmental work
(e.g., infant controlled habituation, gaze-contingent presentation) as well as the fact
that violations of hyperbolic discounting with brief (<1 sec) delays have occasionally
been reported (e.g., Cerekwicki & Grant, 1967; Cerekwicki, Kantowitz, & Grant, 1969;
Landauer, 1964), our findings may be instructive for developmental researchers
employing contingencies in studies of infants of other species, including humans.
GENERAL METHOD
Ethics statement
All animal care and procedures were conducted in strict adherence to the guidelines of
the National Institutes of Health Guide for the Care and Use of Laboratory Animals
and were approved by the Institutional Animal Care and Use Committee of Florida
International University.
Subjects
Fertile, unincubated eggs were received weekly from a commercial supplier (Strickland,
Pooler, GA, USA) and incubated in a Grumbach BSS 160 Incubator (Munich, Ger-
many), maintained at 37.5°C and 70% relative humidity. Twenty-four hours prior to
hatch, embryos were transferred to a Grumbach S84 Hatcher, maintained at 37.5°C
and 80% relative humidity. Shortly after hatch, chicks were transferred to a sound-
proofed rearing room and placed in groups of 10–15 same-aged chicks, to mimic
brood conditions typical for bobwhites (Stokes, 1967). These groups were housed in
plastic tubs (25 cm wide 915 cm high 945 cm long) on shelves in a Nuaire Model
NU-605-500 Animal Isolator (Plymouth, MN, USA). Ambient air temperature was
maintained at approximately 35°C(.5°C each day post-hatch) in the rearing room
and between 30 and 32.8°C in the room where training and testing sessions took place.
Except during training and testing sessions, food and water were available ad libitum.
A range of 18–23 chicks were trained and tested at each delay, within each condition
of the study, drawn from two or more weekly batches to minimize the influence of any
inter-batch variability on the results.
Apparatus
All training and testing sessions were conducted in a large circular arena
(dia =130 cm, height =24 cm) within a sound-attenuated room, non-adjacent to the
rearing room (see Figure 1). The surface of the arena was constructed of plywood,
painted flat black. The walls were constructed of sheet metal, formed into a circle, cov-
ered by a layer of sound-attenuating foam, covered by a layer of opaque black cloth.
Loudspeakers were hidden on opposite sides of the arena, beneath this cloth covering,
at chick level. These were wired to independent RCA SA-155 amplifiers, each con-
nected both to a Sony CDP-XE370 CD player (used during testing) and an isolated
RCA output channel of an M-Audio Audiophile 2496 Sound Card (used during
training). A video camera, mounted on the ceiling above the arena, and a microphone,
placed beneath the center of the arena, gave experimenters visual and auditory access
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
4HARSHAW &LICKLITER

to all sessions. Prior to each session, sound pressure levels at the start location for
chicks placed in the arena (a point equidistant from both speakers on the periphery of
the arena) were calibrated to a maximum of 65 dB for both speakers using a Br€
uel &
Kjaer Model 2232 sound-level meter (B & K Instruments, Marlborough, MA, USA).
At the start of each session, a single chick was placed in an opaque plastic start box at
the start location and left for a period of 30–60 sec of adjustment prior to the begin-
ning of stimulation and/or data collection. All stimulus deliveries and behavioral
observations were made using custom Visual Basic programs.
Auditory stimuli
Two variants of a bobwhite maternal assembly call, Calls A and B (Heaton, Miller, &
Goodwin, 1978), cleaned of background noise by The Borror Laboratory of Bioacous-
tics (Columbus, OH), were used. Both are similar in phrasing, repetition rate and fre-
quency modulation. Call A was recorded in the context of a bobwhite hen leading her
young away from the nest, whereas Call B was given by a bobwhite hen in a colony in
which no chicks were present (Heaton et al., 1978). Despite this, a number of studies
have shown that bobwhite chicks have no naive preference between these calls and are
fully capable of forming a preference for either call given sufficient exposure (Harshaw
& Lickliter, 2007; Honeycutt & Lickliter, 2001, 2002; Lickliter & Hellewell, 1992). Nev-
ertheless, chicks are generally more responsive to Call A than B (i.e., form a stronger
preference given equivalent exposure; Harshaw & Lickliter, 2007), so Call A was used
as the training stimulus throughout the current study.
Figure 1 Experimental setup. Layout of the arena used for training and testing, with the location of
speakers and the start area for chicks indicated. Speakers were hidden behind the cloth covering of
the arena, at chick-level, and the arena was devoid of distinguishing features and landmarks.
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
TIMING-SPECIFIC DEFICIT IN AUDITORY LEARNING 5

Training sessions and delays
All training sessions were conducted on individual chicks on day one, two, or three
post-hatch (day zero being the day of hatch), consisting of 5-min contingent exposure
to bobwhite maternal Call A (Heaton et al., 1978). Sessions were always exactly 5 min
long—an interval that was independent of the number of vocalizations/stimulus pre-
sentations occurred. Presentation of the call was made contingent on chick contact
calls (Harshaw & Lickliter, 2011; Harshaw et al., 2008), which are easily distinguish-
able from other chick vocalizations, composed of a rapid string of high frequency
“peeps.” Bobwhite chicks typically begin contact calling shortly after separation from
conspecifics and/or placement in the training/testing arena, making them an easy
behavior to employ as an operant. Being a ground-dwelling bird, bobwhites are fre-
quently out of visual contact with conspecifics and bobwhite chicks use contact calls to
maintain near constant contact with their mothers and other chicks in the wild
(Stokes, 1967). Although chicks will sometimes vocalize simultaneously, they tend to
“take turns, as the onset of an auditory stimulus shortly after their own call tends to
temporarily inhibit further vocalization and elicits approach (Harshaw & Lickliter,
2007). These ecological considerations make the chick contact call operant and mater-
nal call response a highly useful behavioral system, particularly for examining ques-
tions regarding the formation of preferences for maternal stimuli.
Prior to training, each chick was placed in an opaque plastic start box for 30–
60 s of acclimation. Upon release from the start box—after which chicks moved
freely about the arena for the duration of the test—an experimenter observed and
tracked chick behavior by clicking a specific button with a computer mouse each
time the chick began vocalizing. The button was released when the chick ceased
vocalizing, triggering a single playback of the maternal call (~3 sec) after a 0, 65,
190, 250, 315, 375, 440, 500, 565, 690, 815, 1065, 1315, 1565, or 2,065 msec nominal
delay. For all conditions except the Random Delay (RD) condition (Experiment 4),
each chick heard the maternal call at a single nominal delay, chosen prior to training.
Except in the RD condition, all delays reported are nominal, given that experimenter
reaction time introduced an additional delay. Chicks could also introduce error by
pausing briefly and adding extra peeps—which happened only on rare occasions—or
by beginning to vocalize again prior to the onset of playback. The latter type of
error seldom occurred with brief delays, increasing in frequency with delays greater
than 1,284 msec. To quantify experimenter-introduced error, five-six training sessions
were recorded from six of the experimenters who performed training. The audio
tracks from these 32 training sessions were digitized and analyzed using Magix
Audio Studio, Ver. 10, for Windows. Obtained delays were quantified for each chick
vocalization/maternal call playback and nominal delays subtracted from these. After
the removal of all instances where chicks began vocalizing again prior to the onset
of maternal call playback, stimulus playbacks had an average experimenter-intro-
duced delay of 200–300 msec (mean =219 37 msec). To simplify presentation, all
delays will refer to nominal delays plus this average error (rounded to 220 msec),
unless otherwise stated.
Although the most common response to separation from the brood is for chicks to
emit contact vocalizations, some chicks freeze under such conditions. Such freezing is
thought to be an adaptive anti-predator response or a sign of high arousal and/or anx-
iety (e.g., Gallup & Suarez, 1980). At the onset of training, experimenters thus played
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
6HARSHAW &LICKLITER

the maternal call up to five times (range =1.44–2.86, across experiments) non-contin-
gently, to coax non-vocal chicks into vocalizing (see Harshaw & Lickliter, 2007). Any
chick that failed to respond after these presentations was considered unresponsive and
removed from the study. To prevent the development of side-biases, the side of call
playback was randomized during all sessions. To prevent chicks from lingering in the
vicinity of a single speaker for long periods of time, the program used for randomiza-
tion and stimulus playback was set to avoid runs greater than two, producing a pat-
tern of semi-random switching of sides from playback-to-playback.
Simultaneous choice testing
All testing sessions took place approximately 24 h following training, on days two,
three or four post-hatch. Testing sessions were identical across conditions and experi-
ments, consisting of a 5-min simultaneous choice test between the familiarized (Call A)
and a novel (Call B) maternal call, during which chicks moved freely about the testing
arena. Both calls were played on continuous loop, at equal repetition rates, from
opposite sides of the arena, for the duration of each test. The calls were counterbal-
anced such that for half of the subjects in each condition Call A was played from one
speaker and in the other half Call B was played from that speaker. A semicircular
approach area, representing approximately 5% of the total surface area of the arena,
was demarcated around each speaker on the monitor used by experimenters for
observing testing sessions. Upon entry into an approach area, the experimenter clicked
on one of two buttons within a custom Visual Basic program. The button was held
down until the chick exited the approach area. This provided frequency tallies of
entries into both areas, cumulative scores for duration of time spent within each area,
and scores for latency of approach to each area.
Data analysis for simultaneous choice tests
Raw duration scores were converted into categorical “preferences”, so that non-para-
metric chi-square tests could be performed on their distributions. Chicks failing to
spend at least 30 sec (10% of the trial) in an approach area were scored as non-
responders and excluded from further analyses. Of the remaining subjects, chicks fail-
ing to spend at least twice as long in one approach area as in the other were scored
as displaying no preference. A chick was thus scored as displaying a preference for a
call only if the chick spent at least 30 sec in the approach area for that call and at
least twice as long in that approach area as in the other. A latency score of 300 sec
and a duration score of zero were assigned for any area not entered by a chick dur-
ing a testing session. Duration and latency scores were converted into proportion of
total duration (PTD) for Call A (Duration
A
/(Duration
A
+Duration
B
)) and propor-
tion of trial to approach (PTTA) scores (Latency/300), respectively. Single-sample t-
tests against the null hypothesis of chance responding were conducted for PTD and
PTTA difference (familiar minus unfamiliar) scores for each sub-condition. All statis-
tics were calculated using R, version 3.1.3 (The R Foundation for Statistical Com-
puting, Vienna, Austria), with a 5% criterion for significance (two-tailed). Curve
fitting and model comparison were performed using GraphPad Prism, version 6
(GraphPad Software, Inc., La Jolla, CA, USA) and fits to the hyperbolic decay
model were evaluated using F-tests for Lack of Fit. Given the large number of con-
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
TIMING-SPECIFIC DEFICIT IN AUDITORY LEARNING 7

ditions run, Bonferroni corrections were employed throughout to ensure a familywise
error-rate of p<.05.
EXPERIMENT 1: DELAY AND LEARNING AT 24 H
As noted at the outset, our first experiment sought to define a temporal envelope for
the perception of contingency in bobwhite quail chicks. That is, at what delay would
chicks no longer respond to an auditory stimulus as if it were related to their own
behavior? Our first experiment addressed this question by examining the effects of
delay on the acquisition of auditory preferences for a species-typical, bobwhite mater-
nal call in day-old bobwhite chicks. A rapid, hyperbolic decline in learning or reward
value with increased delay between a behavior and its consequence is the typical find-
ing in operant paradigms (cf. Catania, 2006). We thus expected to find a simple, hyper-
bolic decline in learning in day old chicks as the delay between chick contact
vocalization and maternal call playback was increased from zero to 2,000 msec.
Method
Subjects
One-day-old Northern bobwhite (Colinus virginianus) chicks (N=492) served as
subjects. Of these, 85 chicks (17.3%) failed to respond (i.e., vocalize) during training,
five chicks (1.2%) were trained but not tested due to scheduling errors, and 103
(20.9%) failed to respond during simultaneous choice testing. The final sample thus
included 300 day-old chicks, distributed between 15 delay conditions (N=18–
23 chicks/delay; mean =20 .5).
Procedure
At 24 h of age, each chick was provided with a single 5 min vocal-auditory training
session, as described in the General Method. At 48 h of age, each chick was tested
using a simultaneous choice between the familiar (Call A) and an unfamiliar (Call B)
bobwhite maternal call (see Harshaw & Lickliter, 2007).
Results and discussion
Chick duration (PTD) and latency to approach the familiar call (PTTA) scores are
shown in Figures 2a and 3a, respectively. We expected to find a negatively accelerated
hyperbolic decline in learning or preference for the familiar call with increased delay,
following the model V=1/(1 +kD), where Vis reward value, Dis delay and kdeter-
mines the rate of decline (Mazur, 1987). The best fit of the hyperbolic decay model to
chicks’ preference (PTD) scores nonetheless provided an inadequate fit (F=3.81,
p<.0001). A series of polynomial regressions (up to 6th order) similarly failed to pro-
vide a fit to the data that was superior to that provided by a linear model.
Instead of a hyperbolic decline in learning, chicks appeared to experience intermit-
tent interference in their ability to learn and remember (i.e., acquire a preference for)
the familiar call, particularly with delays between 450 and 900 msec (see Table 1). This
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
8HARSHAW &LICKLITER

result was surprising given the ease with which bobwhite chicks generally form strong
preferences for the particular training stimulus used, maternal Call A (see Harshaw &
Lickliter, 2007). Nevertheless, exceptions to strict hyperbolic discounting have been
noted (e.g., Cerekwicki & Grant, 1967; Cerekwicki et al., 1969; Landauer, 1964), particu-
larly for “very brief” delays of less than a second (see Schneider, 1990). Studies of brief
unsignaled delay of reinforcement have, for example, often reported an increase in
response rate with delayed compared to immediate reinforcement (Arbuckle & Lattal,
1988; Lattal & Ziegler, 1982; Richards, 1981; Sizemore & Lattal, 1978). These studies
suggest that optimal learning in contingency-based or operant paradigms may, under
some circumstances, have an optimal delay, similar to the optimal inter stimulus interval
in Pavlovian paradigms (see Jones, 1962; Rescorla, 1988). Although it was theoretically
possible that we had inadvertently created conditions similar to those in other studies
that have found violations of strict hyperbolic decay, we sought to replicate our findings
with older chicks, to verify that our result was not simply the result of the noisiness of
simultaneous choice data obtained from very young (i.e., day-old) birds.
Figure 2 11
Mean proportion of total duration for the familiar call (PTD; i.e., duration of time spent
in proximity to the familiar relative to both maternal calls) during simultaneous choice testing
(SEM) as a function of delay between offset of chick vocalization and call playback. A score of .5
indicates chance responding. PTD scores for chicks trained at (a) 24 (b) 48 and (c) 72 h of age
(N=300, 301, and 267, respectively). The box demarcates analyses of PTD data pooled across all of
these ages, including both (d) average scores and (e) measures of consistency across these ages.
Specifically, sliding window correlations presented in (e) are an average of three-delay sliding windows
(Schulz & Huston, 2002) comparing PTD scores at 24–48, 24–72, and 48–72 h. In addition, (f)
displays scores for day-old chicks trained with a single speaker (N=287) compared to those trained
with two speakers (i.e., non-switching vs. switching of spatial locations).
Colour online, B&W in print LOW RESOLUTION FIG
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
TIMING-SPECIFIC DEFICIT IN AUDITORY LEARNING 9

EXPERIMENT 2: DELAY AND LEARNING AT 48 AND 72 H
Our second experiment examined the effects of delay on the acquisition of preferences
for a contingently presented maternal call in 2- and 3-day-old bobwhite quail chicks.
As in the first experiment, we expected to find a negatively accelerated hyperbolic
decline in learning with increased delay between chick contact vocalizing and contin-
gent presentation of the bobwhite maternal call.
Method
Subjects
Two- and three-day-old Northern bobwhite chicks (N=1,078) served as subjects.
Of these, 293 chicks (27.2%) failed to respond (i.e., vocalize) during training, 17
chicks (2.2%) were trained but not tested, due to scheduling errors, and 200 (18.6%)
failed to respond during simultaneous choice testing. The final sample thus included
Figure 3 12
Latency to approach the familiar call. Mean (SEM) proportion of trial elapsed prior to
approach (PTTA
familiar
PTTA
unfamiliar
), as a function of the delay between the offset of chick
contact vocalization and onset of playback. A score of zero indicates chance responding or no
preferential discrimination of the calls, whereas scores below zero indicate shorter latencies for the
familiar call. PTTA difference scores for chicks trained at (a) 24 (b) 48 and (c) 72 h of age (N=300,
301, and 267, respectively), in addition to (d) average scores, across these ages and (e) scores for day-
old chicks trained with a single speaker (N=287) compared to those trained with two speakers
(Experiment 3). 3
Colour online, B&W in print LOW RESOLUTION FIG
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
10 HARSHAW &LICKLITER

301 2-day-old and 267 3-day-old chicks, distributed between 15 delay conditions
(N=18–24 chicks/delay; 48 h mean =20.1 .95; 72 h mean =19.1 .6).
Procedure
At 48 or 72 h of age, each chick was provided with a single 5 min vocal-auditory
training session, as described in the General Method. Two-day-old chicks were trained
at all 15 of the delays used in Experiment 1, whereas 3-day-old chicks were trained at
only 14 of these delays—the last delay being excluded because of diminishing returns
and the increased difficulty of working with 2-to-4-day-old chicks. At 72 and 96 h of
age, respectively, each chick was tested using a simultaneous choice between the famil-
iar (Call A) and an unfamiliar (Call B) bobwhite maternal call (see Harshaw & Lick-
liter, 2007).
Results and discussion
Despite some differences, we found a relatively high degree of consistency in chick per-
formance across ages. Chick duration (PTD) and latency to approach the familiar call
(PTTA) scores are shown in Figures 2b,c and 3b,c (48 h/72 h), respectively. As can be
seen, chicks trained at 48 h of age showed interference in their learning of the familiar
call similar to that shown by day-old chicks (PTD: r=.52, p=.06; PTTA: r=.27,
p=.34). That is, 2-day old chicks appeared to have greater difficulty learning and
remembering the call when trained with delays between 450 and 900 msec (see
Table 2). As in Experiment 1, the best fit of the PTD data from 2-day-old chicks to
TABLE 1
Preferences Displayed by Chicks Trained at 24 h of Age
Delay (in ms) n responding
Call preference Chi-Square tests
ABNP v
2
p-Value E.S. (W)
220 18 15 1 2 20.3 .00004 1.06
285 23 16 2 5 14.2 .00084 .79
410 19 19 0 0 38.0 .00000 1.41
470 18 13 5 0 14.3 .00077 .89
535* 22 12 5 5 4.5 .10782* .45
595 22 18 2 2 23.3 .00001 1.03
660* 18 8 4 6 1.3 .51342* .27
720 21 15 3 3 13.7 .00105 .81
785 21 16 2 3 17.4 .00016 .91
910* 19 12 2 5 8.3 .01564* .66
1,035 18 15 1 2 20.3 .00004 1.06
1,285 22 17 2 3 19.2 .00007 .93
1,535 21 16 4 1 18.0 .00012 .93
1,785* 19 12 4 3 7.7 .02145* .64
2,285* 19 12 3 4 7.7 .02145* .64
Note. E.S., Effect size.
Chi square tests are against the null distribution of preferences (i.e., 1/3, 1/3, 1/3) between categories.
*Delay conditions with non-significant chi-squared values (p≥.0033 [a=.05/15]).
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
TIMING-SPECIFIC DEFICIT IN AUDITORY LEARNING 11

the hyperbolic decay model provided an inadequate fit (F=3.81, p<.0001). As can
be seen in Figures 2c and 3c, chicks trained at 72 h of age showed PTD and PTTA
difference scores that were less similar to chicks trained at 24 h of age (r=.39,
p=.189, and r=.01, p=.652, respectively) than those of chicks trained at 48 h.
The overall pattern of interference in chick preferences for the familiar call as a func-
tion of delay was nonetheless similar, with chicks, again, showing greater interference
in the learning and memory for the familiar call with delays in the range of 450–
900 msec (see Table 3).
Chick preferences for the familiar call as a function of delay were thus fairly consis-
tent over the 24–72 h age range (see Figures 2d and 3d, depicting averaged PTD and
PTTA scores, respectively). This consistency was explored further via a series of sliding
window correlations (see Schulz & Huston, 2002) on PTD and PTTA scores, with win-
dows of three time points (i.e., delays).
1
As can be seen in Figures 2e and 3e, which
depict both the average sliding window correlation for all possible age comparisons
and the average variance (i.e., standard error) across ages for PTD and PTTA scores,
respectively, the greatest coherence in performance between-ages occurred with delays
in the range of 700–1,400 msec. Additionally, plots of standard errors averaged across
ages are shown. Analysis of Rhythmic Variance (ANORVA) is based on the straight-
forward assumption that for any rhythmic process, samples at or near peaks and
troughs of a cycle will have lower variance than samples at other points in a cycle
(Celec, 2004). In our data, the lowest variance in PTD scores occurred with 410 and
TABLE 2
Preferences Displayed by Chicks Trained at 48 h of Age
Delay (in ms) n responding
Call preference Chi-Square tests
ABNP v
2
p-Value E.S. (W)
220 20 16 2 2 19.6 .00006 .99
285 19 13 0 6 13.4 .00125 .84
410 18 13 1 4 13.0 .00150 .85
470* 23 13 3 7 6.6 .03672* .54
535 21 14 0 7 14.0 .00091 .82
595* 18 11 3 4 6.3 .04214* .59
660* 24 15 6 3 9.8 .00764* .64
720* 19 10 4 5 3.3 .19562* .41
785* 19 9 5 4 1.7 .43080* .36
910* 20 13 4 3 9.1 .01057* .67
1,035 20 18 1 1 28.9 .00000 1.20
1,285* 21 14 3 4 10.6 .00506* .71
1,535 21 17 2 2 21.4 .00002 1.01
1,785* 19 9 5 5 1.7 .43080* .30
2,285* 19 10 5 4 3.3 .19562* .41
Note. E.S., Effect size.
Chi square tests are against the null distribution of preferences (i.e., 1/3, 1/3, 1/3) between categories.
*Delay conditions with non-significant chi-squared values (p≥.0033 [a=.05/15]).
1
Although this procedure was not ideal given the uneven spacing of delays, these analyses give a sense of
the between-age temporal coherence indicated by our data.
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
12 HARSHAW &LICKLITER

1,035 msec delays, suggesting that these are near the peaks and/or troughs of an
underlying rhythmic process (see Celec, 2004).
In summary, the observed timing-specific deficit or “blink” in auditory learning
with delays in the range of 450–900 msec was found to be relatively stable over the
first few days post-hatch. However, due to some apparent developmental shift, affect-
ing the responding of 3-day old chicks to 400–700 msec delays, the greatest overall
coherence between ages was found for delays of 700–1,400 msec, with peak learning
consistently occurring with delays of ~1,000 msec. Although there are many potential
causes of this phenomenon, one strong possibility was that chicks were anticipating
the location of the maternal call from trial to trial and preemptively deploying their
attention prior to each playback, resulting in interference with learning, time-locked
to this attentional shift. Two facts suggested this possibility. First, chicks are highly
motivated to approach the maternal call—typically making a “bee line” for the
active speaker at the onset of each playback. Second, the fact that playback location
was inherently uncertain (i.e., semi-random, with no runs greater than two) meant
that there was a high likelihood for chick expectation to conflict with playback loca-
tion. In fact, if chicks were employing the simple “strategy” of expecting playback
from the same location as that of the immediately prior playback then they would
err on roughly 2/3 of the trials, as only 1/3 of trials, on average, involved an imme-
diate repetition of playback location. The additional time needed to correct for any
such errors—localizing and orienting/locomoting toward the maternal call—on trials
involving a switch of spatial location may have thus significantly interfered with
chick learning on such trials; again, in a manner time-locked to the initial shifting of
attention to the incorrect location. Our third experiment was designed to explore this
hypothesis.
TABLE 3
Preferences Displayed by Chicks Trained at 72 h of Age
Delay (in ms) n responding
Call preference Chi-Square tests
ABNP v
2
p-Value E.S. (W)
220 18 14 1 3 16.3 .00028 .95
285 19 13 1 5 11.8 .00275 .79
410* 19 9 3 7 3.0 .22908* .39
470 18 13 1 4 13.0 .00150 .85
535* 19 12 4 3 7.7 .02145* .64
595* 18 9 3 6 3.0 .22313* .41
660 20 15 2 3 15.7 .00039 .89
720 20 14 4 2 12.4 .00203 .79
785* 18 12 3 3 9.0 .01111* .71
910* 19 13 4 2 10.8 .00442* .76
1,035 21 15 0 6 16.3 .00029 .88
1,285* 20 11 4 5 4.3 .11648* .46
1,535* 18 11 4 3 6.3 .04214* .59
1,785 20 14 3 3 12.1 .00236 .78
Note. E.S., Effect size.
Chi square tests are against the null distribution of preferences (i.e., 1/3, 1/3, 1/3) between categories.
*Delay conditions with non-significant chi-squared values (p≥.0036 [a=.05/14]).
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
TIMING-SPECIFIC DEFICIT IN AUDITORY LEARNING 13

EXPERIMENT 3: SPATIAL SWITCHING AND LEARNING
To test the hypothesis that the switching of location between playbacks resulted in the
observed timing-specific deficit or “blink” in learning, we provided day-old chicks
training in which the location of playback was chosen randomly prior to the onset and
fixed throughout training. We predicted that if the timing-specific deficit in learning
were the result of difficulties tied to the temporal dynamics of attentional shifting in a
novel and uncertain environment, then day-old chicks trained with only a single spatial
location would fail to show the previously documented deficit in learning with delays
of 450–900 msec.
Method
Subjects
Day-old Northern bobwhite chicks (N=428) served as subjects. Of these, 59 chicks
(13.8%) failed to respond (i.e., vocalize) during training, 19 chicks (5.1%) were trained
but not tested due to scheduling errors, and 63 (21.9% of chicks tested) failed to
respond during testing. The final sample thus included 287 day-old quail chicks, dis-
tributed between 14 delay conditions (N=18–23 chicks/delay; mean =20.5 1.1).
Procedure
Single-speaker training sessions were identical to the vocal training sessions of
Experiments 1 and 2, except that the side of call playback was chosen randomly by a
computer program at the onset of each session and remained fixed throughout. The
fixing of playback location eliminated the need for chicks to re-orient following each
vocalization and playback.
Results and discussion
Chicks trained with only a single playback location did not show as pronounced of a
deficit specific to delays of 450–900 msec as was exhibited by chicks trained with two
playback locations, in Experiments 1 and 2. As can be seen in Table 4, which displays
tallies of chick preferences, the only delay condition in which chicks trained with a sin-
gle location showed reduced learning of the familiar call were the 910 and 1,785 msec
conditions, in contrast to Experiment 1, in which chicks showed poor learning of the
call at 535, 660 and 910 msec, in addition to the longer 1,785 and 2,285 msec, delays.
Comparing chick duration scores for the familiar call (i.e., PTD; see Figure 2f), there
was relatively little correlation between the average PTD scores of chicks provided
with training with one vs. with two locations (r=.21, p=.49), indicating that chicks
responded quite differently to training with a single speaker. For example, the delays
at which day-old chicks showed the poorest performance when trained with two loca-
tions (i.e., 535 and 660 msec) were the specific delays with the highest performance in
single-location training (comparison: t=1.15, p=.132, and t=2.62, p<.008, respec-
tively). Nevertheless, as in the two-speaker training, a hyperbolic model provided an
inadequate fit to the data (F=2.17, p<.02). Beyond this, there was a trend toward a
more linear decline in learning with delay for chicks trained with a single speaker
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
14 HARSHAW &LICKLITER

(r=.43, p=.139) than for same-age chicks trained with two speakers in Experiment
1(r=.14, p=.645). In contrast to PTD scores, average latency (i.e., PTTA differ-
ence; see Figure 3f) scores for chicks trained with a single speaker tended to be posi-
tively correlated with PTTA those of chicks trained with two speakers (r=.24,
p=.42). Nevertheless, a comparison of average scores for all delays within the 450–
900 msec windows across conditions (see Figure 4) revealed significantly improved
PTD scores for the familiar (t=2.01, p<.05) and a similar trend, toward shorter
latency scores for the familiar (t=1.08, p=.28), in chicks that underwent training
with a single location compared to two locations.
These results indicate that the switching of playback location from trial to trial
likely interfered with chicks’ learning the familiar call in our initial experiments when
playback occurred 450–900 msec after chick vocalizations. Chicks given training that
was identical to that used in Experiments 1 and 2 with the exception that playback
occurred from a single, randomly chosen spatial location appeared to have less diffi-
culty learning and remembering the call with 450–900 msec delays (see Figure 4). They
nonetheless still showed some difficulty with these delays relative to an ~1,000 msec
delay. One possible explanation for this result was that chicks are highly active (i.e.,
motorically) during training, including periods immediately following call playback.
Chicks may have thus drifted from the speaker location—becoming “lost,” given the
uniform environment and lack of landmarks—and thus still have been making atten-
tional errors, requiring reorienting to the auditory stimulus from trial to trial when
trained with a single playback location. We thus sought to obtain finer-grained data
on chick behavior as a function of both delay and trial type (i.e., trials involving play-
back form the same location vs. a switch of spatial location) during training to shed
light on the spatiotemporal factors influencing chick learning in our contingent play-
back paradigm.
TABLE 4
Preferences Displayed by Chicks Trained With a Single Location at 24 h
Delay (in ms) n responding
Call preference Chi-Square tests
ABNP v
2
p-Value E.S. (W)
220 23 16 1 6 15.2 .00050 .81
285 19 14 2 3 14.0 .00091 .86
410 18 16 2 0 25.3 .00000 1.19
470 23 18 1 4 21.5 .00002 .97
535 20 14 1 5 13.3 .00129 .82
595 18 15 2 1 20.3 .00004 1.06
660 19 16 0 3 22.8 .00001 1.10
720 23 16 1 6 15.2 .00050 .81
785 20 15 3 2 15.7 .00039 .89
910* 23 14 3 6 8.4 .01474* .61
1,035 22 16 3 3 15.4 .00046 .84
1,285 20 16 1 3 19.9 .00005 1.00
1,535 19 14 2 3 14.0 .00091 .86
1,785* 20 13 3 4 9.1 .01057* .68
Note. E.S., Effect size.
Chi square tests are against the null distribution of preferences (i.e., 1/3, 1/3, 1/3) between categories.
*Delay conditions with non-significant chi-squared values (p≥.0036 [a=.05/14]).
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
TIMING-SPECIFIC DEFICIT IN AUDITORY LEARNING 15

EXPERIMENT 4: RANDOM DELAYS AND SPATIAL SWITCHING
Our working hypothesis was that chicks preemptively deploy their attention to an “ex-
pected” location prior to each playback and that mismatch between chick expectations
and actual playback location caused chicks difficulty specifically when playback
occurred 450–900 msec following chick vocalization. To further explore the mecha-
nisms underlying this timing-specific deficit, we provided day-old chicks with 5-min ses-
sions in which they received playback of the maternal call at a random delay between
zero and 1250 msec each time that they vocalized. These sessions were videotaped and
the obtained delay (i.e., actual rather than nominal delay) quantified for each playback
or “trial.” Additionally, each trial was scored for a number of chick behaviors related
to orienting and responsivity to the maternal auditory stimulus. Trials were analyzed
further by partitioning them into those involving playback from the same location as
that of the previous playback (“Same” trials) and those involving a switch of spatial
location (“Switch” trials). Based on the hypothesis that chicks allocate their attention,
in an obligatory manner, to same locations after vocalizing, we predicted that chicks
would show an advantage orienting and navigating toward the auditory stimulus on
Same compared to Switch trials. We also expected that a comparison of fluctuations in
Figure 4 13
Single vs. two-location training with 450–900 msec delays at 24 h of age. Bars indicate
average PTD and PTTA difference scores across all delay conditions within the 450–950 msec range.
Higher PTD scores indicate more time spent in proximity to the familiar call (.5 being chance
responding) whereas lower PTTA difference scores indicate shorter latencies to approach the familiar
call (zero being chance responding). Significance reported is for independent sample t-tests. Note that
the 910 msec delay condition was purposefully excluded from these analyses. If included, the same
trends in the data still hold for both PTD (t=1.72, p=.087) and PTTA difference (t=1.06,
p=.292) scores.
LOW RESOLUTION FIG
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
16 HARSHAW &LICKLITER

chick responsivity as a function of delay on these types of trial would provide insight
into the influence of attention and orienting on chick learning of the maternal call.
Method
Subjects
Twenty-two day-old bobwhite chicks participated in the experiment. Of the 22
chicks trained, four were excluded from statistical analyses for not approaching the
speaker emitting the call on at least 60% of trials. The final sample thus consisted of
18 day-old chicks.
Procedure
Chicks were provided with 5-min training sessions that were identical to those
employed in Experiments 1 and 2 except that each playback of the bobwhite maternal
call occurred at a delay between zero and 1,250 msec that was chosen randomly by a
computer program prior to each playback.
Data analysis
Training sessions were videotaped and these videos later scored for a variety of
behaviors related to orienting, approach, and responsivity to the auditory stimulus.
The audio tracks from these videos were also digitized and analyzed using Magix
Audio Studio, Version 10, for Windows for the purpose of measuring obtained delays
between each chick vocalization and maternal call playback. Trials were defined as a
single chick vocalization followed by a single maternal call playback, ending upon the
onset of the next chick vocalization. For each trial, several behaviors were scored as
Boolean variables, including (1) whether or not the chick turned toward the correct
location or showed any other evidence of orienting to the stimulus (e.g., stopping and
listening), (2) whether or not the chick took at least one step in the direction of (i.e.,
approached) the correct location, and (3) whether or not the chick entered the
approach area (identical to approach areas used in simultaneous choice testing; i.e.,
5% of the surface area of the arena) surrounding the correct location. This allowed
for the easy calculation of the probability of each behavior as a function of delay and
trial type—i.e., Pr(orient), Pr(approach), Pr(enter). In addition, (4) the total number of
entries to the correct location and (e) duration of time spent in proximity to the cor-
rect location were scored for each trial. Given that chicks often sprinted in what
looked, to the naked eye, to be an apparently random direction upon the onset of
playback, each chick’s starting location and initial angle of approach relative to the
correct location were also quantified, providing a measure of the chick’s accuracy in
orienting to the stimulus on each trial.
Given that the goal of the experiment was to examine chick orienting and approach
behavior, only trials following the first successful approach and entry into a correct
approach area were considered. Except for sequential analyses, only trials in which
chicks were outside of the correct approach area at the onset of playback—in which
chicks needed to orient and approach—were analyzed. After removing all such invalid
trials (N=30), there were a total of 347 usable trials acquired from the 18 chicks that
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
TIMING-SPECIFIC DEFICIT IN AUDITORY LEARNING 17

made up the final sample. Overall differences between Same and Switch trials were
tested using paired-sample ttests. In addition, data were averaged into 50 msec time
bins (based on obtained rather than nominal delays) and sliding window correlations
performed on the resulting time series for Same and Switch trials (see Schulz & Hus-
ton, 2002). All tests were performed using NCSS 2007 for Windows (J. Hintze, Kays-
ville, Utah), evaluated using an aof .05.
Results and discussion
Confirming our prediction, chicks performed significantly better on Same than Switch
trials on several behavioral measures of responsivity and orienting (see Figure 5). First,
analysis of chicks’ initial angle of approach showed that chicks were significantly better
Figure 5 14
Chick orienting and approach during training. Measures of chick responsivity to the
maternal call during training involving playback at random (0–1250 msec) delays and two spatial
locations, separated into trials involving playback from the same location as that of the previous
playback (same trials) or a switch of location (switch trials). Mean (SEM) (a) duration of time spent
in proximity to and activity near the call, (b) proportion of trials chicks entered, approached and
failed to orient toward the correct approach area, and (c) initial headings, expressed as absolute
values of deviations from zero degrees. Chicks generally sprint rapidly toward the speaker emitting
the call at the onset of playback, but frequently arrive at the correct location only after making a
series of corrections to their trajectories. The initial heading of a chick thus conveys information
about how well the chick has localized the auditory stimulus prior to the onset of approach.
Colour online, B&W in print LOW RESOLUTION FIG
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
18 HARSHAW &LICKLITER

at orienting to the correct speaker on Same (mean =45.3°) compared to Switch
(mean =63.9°) trials (F=8.45, p<.005; Figure 5c). As can be seen in Figures 5a,b,
chicks also showed a tendency to approach the correct speaker on a greater proportion
of Same than Switch trials (t=1.26, p=.104) and entered the correct approach area
on a significantly greater proportion of Same than Switch trials (t=1.95, p=.026).
In addition, chicks spent significantly more time in proximity to and showed signifi-
cantly greater activity levels near the correct speaker on Same compared to Switch tri-
als (t=4.07, p<.00003, and t=2.66, p<.005, respectively).
Although these results provide some support for the idea that chicks deploy their
attention to Same locations prior to the onset of playback, another explanation would
be that chicks simply happen to be closer to the correct location and thus face an
easier task on Same relative to Switch trials. A comparison of chick start locations
showed that chicks were indeed nearer to the correct speaker at the onset of Same
(72.6 5.4 cm) compared to Switch (88.3 3.6 cm) trials (t=4.76, p<.000003) at
the onset of playback. Nonetheless, an analysis of chick responsivity on the two trial
types as a function of delay suggests that mere proximity cannot fully explain the per-
formance advantage displayed by chicks on Same trials. That is, chicks displayed a
pattern of rapidly fluctuating responsivity to the auditory stimulus on both types of
trial (see Figure 6a). Specifically, on Same trials chicks showed periods of high respon-
sivity alternating with periods of total unresponsiveness, suggesting oscillation between
facilitation of performance and inhibition of return (see M€
uller & Kleinschmidt, 2007).
In contrast, chicks showed less extreme fluctuations in responsivity on Switch trials.
Sliding window correlations on the two time series (Schulz & Huston, 2002) revealed
that fluctuations in chick responsivity on the two trial types were initially in phase (or
similar) from zero to 300 msec, became uncoupled (dissimilar) between 400 and
800 msec, and were again in phase from 900 to 1,100 msec (Figure 6a). This result
indicates that the greatest effect of trial type on fluctuations in chick responsivity dur-
ing training—or the time period when trial type matters most—occurs when the audi-
tory stimulus falls 400–800 msec after chicks vocalize. Although this range is not
identical to the 450–900 msec window for disruption of learning found in our initial
experiments, the substantial overlap, combined with the fact that delay was measured
with far greater precision in the present experiment, suggests that chick learning, as
measured in Experiments 1–3, is closely tied to orienting and approach behaviors
exhibited by chicks during training.
Several sources of evidence support this contention. First, we performed a series of
lagged analyses on the same data set described above. We were interested in sequential,
trial to trial dependencies and whether such dependencies would vary as a function of
delay. That is, would delays of 400–800 msec have different downstream consequences
on chick performance than delays outside of this window? We reasoned that if chicks
have particular difficulty localizing stimuli encountered at 400–800 msec delays, then
this should have reliable effects on their performance on subsequent trials, particularly
given that chicks have little choice but to navigate via path integration (i.e., dead reck-
oning) in our training apparatus, which is entirely devoid of landmarks. In particular,
we hypothesized that chicks would be more disoriented and thus perform significantly
worse on trials immediately following those in which playback occurred at 400–
800 msec delays.
Although this simple prediction was not confirmed, a clear pattern of inter-trial
dependency was nonetheless apparent, as can be seen in Figure 6b, which shows
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
TIMING-SPECIFIC DEFICIT IN AUDITORY LEARNING 19

250 msec sliding window correlations between time-lagged performance and Lag 0 per-
formance. Rather than simply performing more poorly on trials following those
involving 400–800 msec delays, a more dynamic pattern was evident: with short (i.e.,
Figure 6 15
Chick responsivity during training as a function of delay and trial type. (a) Proportion of
trials chick entered the correct approach area as a function of the delay between offset of chick
vocalization and initiation of playback on Same and Switch trials (left axis), averaged into 50 msec time
bins. Sliding window correlations are superimposed over these (right axis), to illustrate the robustness
of the pattern obtained to variations in window size (250–450 msec). The “weighted 250 msec”
correlation is an average of 150 msec and 250 msec windows. Values above the dashed line indicate
positive correlations whereas values below the line indicate negative correlations. (b) An analysis of
trial-to-trial dependency of performance during training, with data averaged into 50 msec bins relative
to four lags (Lag 0 being the original data set), and 250 msec sliding window correlations between
these. This figure illustrates the impact on a given trial (lag 0) of having been preceded, at various lags,
by a trial of a particular delay, irrespective of whether the trial was a Same or Switch trial.
Colour online, B&W in print LOW RESOLUTION FIG
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
20 HARSHAW &LICKLITER

0–200 msec) delays, there was strong negative inter-trial correlation in performance.
That is, if chicks performed relatively well on one trial they tended to perform poorly
on subsequent trials, and vice versa, irrespective of the delays encountered on those tri-
als. This was followed by an increasingly positive inter-trial correlation, particularly at
400–800 msec delays, indicating that at these delays and at these delays only, if chicks
performed well on one trial they tended to perform well on subsequent trials, and vice
versa. This was followed again by negative inter-trial correlation for delays greater
than 800 msec. Interestingly, this function appeared to follow a course exactly inverse
to that obtained in our analysis of fluctuations in chick performance on Same and
Switch trials (Figure 6a), and lagged performance was indeed correlated with Same-
Switch correlation, both for Lag 1 (r=.5, F=6.99, p=.0152) and Lag 2 (r=.83,
F=45.1, p<.000001) trials. This relationship decayed, however, by Lag 3 (r=.16,
F=.42, p=.479). This result suggests both that chicks integrate information about
playback location over the course of multiple trials, within a window of at least 20–
30 sec (ITI =11.49 .21 sec), and that playbacks become strong attractors of chicks’
downstream attention only when they fall outside of (before or after) the 400–800 msec
“blink” window.
A parsimonious account of these findings can be provided by assuming that they
are the product of the interplay between two attentional systems well-described in the
human literature: an exogenous or stimulus-driven system—responsible for reflexive,
automatic orienting—and an endogenous or internally driven system, producing “vol-
untary” shifts of attention based on ongoing goals and task-demands (Corbetta &
Shulman, 2002). The exogenous system produces brief facilitation of processing from
an attended or cued location, followed by relatively protracted inhibition of return
(IOR) to that location (Posner & Cohen, 1984)—a pattern widely thought to facilitate
search and foraging (e.g., Bays & Husain, 2012; Klein, 2000; Robertson, Watamura, &
Wilbourn, 2012). The endogenous system has a slower time-course, generating maximal
facilitation of processing 200–300 msec following a cue or shift of attention and does
not generate IOR (Weichselgartner & Sperling, 1987).
Such an approach to analyzing our paradigm would predict that on Same trials the
poorest performance should occur when IOR (to Same locations) is maximal, several
hundred ms after chicks cease vocalizing and shift their attention to the expected loca-
tion. In contrast, on Switch trials the poorest performance should occur during the
first few hundred ms following the offset of chick vocalization, when attentional cap-
ture by the Same location—the location that would be incorrect on ~2/3 trials—should
be maximal. As can be seen in Figure 7, which shows the data for Same and Switch
trials collapsed into 350 msec bins, such a pattern is evident, particularly for measures
of orienting and approach. Our lagged analyses reinforce the contention that reflexive
orienting and IOR are operative above and below the 400–800 msec window, as good
performance outside of this window tends to be followed by poorer performance (and
vice versa)—implying that, in this range, a well-localized stimulus becomes a potent
attractor of attention, resulting in interference on downstream trials (Figure 6b). The
fact that this does not occur during the 400–800 msec window implies the working of
a second process. If the endogenous attentional system were at play, the prediction
would be superior performance on Switch trials during precisely the 400–800 msec win-
dow. Yet, birds exhibit only fair (not facilitated) performance during training (Fig-
ure 7) and impaired learning with delays in this range (see Figures 2 and 3).
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
TIMING-SPECIFIC DEFICIT IN AUDITORY LEARNING 21

Two factors may contribute to this outcome. First, unlike typical search tasks,
which involve exploration of a stimulus array, there are only two locations at which a
stimulus can occur in the present paradigm. Moreover, there are no landmarks that
chicks can use to anchor their search from playback to playback. Our task thus
involves a special (and complex) case of search. In addition, it is possible that the
endogenous attentional system requires strong expectancies and that there is little
room for facilitation via endogenous mechanisms to occur in our paradigm given that
there is no cueing of location and 2/3 of all trials involve a switch of spatial location.
These two possibilities are not mutually exclusive and further experimentation will be
needed to obtain a full understanding of the phenomenon documented here.
GENERAL DISCUSSION
Any study employing contingencies must settle upon values for a few critical parame-
ters of the experiment, one of the most critical being the delay or degree of temporal
contiguity between operant and response (see Watson, 2001; Williams, 2001). Delay
value can be chosen arbitrarily, varied systematically, or if social interaction is the
object of study, allowed to vary naturalistically. The findings presented here illustrate
that this seemingly simple choice can nonetheless be fraught, even in a relatively simple
experimental paradigm. Specifically, we studied the formation of filial auditory prefer-
ences in bobwhite quail chicks in a situation in which chicks vocalized to hear operant
Figure 7 16
Same and Switch trial performance over time. Measures of Lag 0 performance on Same
and Switch trials during training, with data averaged into 350 msec bins. (a) Pr(~Orient) =Proportion
of trials chicks failed to orient (i.e., showed no evidence of orienting) to the correct location. (b) Pr
(Approach) =Proportion of trials chicks took at least one step toward the correct location. (c) Pr
(Enter) =Proportion of trials chicks entered the correct approach area. (d) Activity in the vicinity of
stimulus (i.e., # of entries into the approach area). (e) Duration of time spent in proximity to stimulus
(i.e., time spent in the approach area). Asterisks indicate a significant difference between values within
trial type (Bonferroni test; *p<.05; **p<.002).
Colour online, B&W in print LOW RESOLUTION FIG
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
22 HARSHAW &LICKLITER

playback of a bobwhite maternal call. The results of our first two experiments demon-
strated that 1- to 3-day-old bobwhite quail chicks show a timing-specific deficit or
“blink” in auditory learning when an auditory stimulus occurs 450–900 msec after
their own contact vocalizations in an operant, delay-of-reinforcement paradigm mod-
eled on the naturalistic context of a young chick attempting to localize its mother by
vocally interacting with her (see Harshaw & Lickliter, 2007, 2011; Harshaw et al.,
2008). Our analyses indicate that a quirk of our experimental design, the semi-random
switching of spatial locations during training (employed to prevent birds from forming
side biases) likely caused the deficit, albeit in interaction with chicks’ developing orient-
ing and attentional abilities.
We initially hypothesized that the timing-specific deficit observed in Experiments 1
and 2 was caused by chicks preemptively deploying their attention to the previous
playback location prior to the onset of each new playback, causing chicks difficulty
given the frequent switching of spatial location in our paradigm. In Experiment 3,
day-old chicks were thus provided with training sessions in which playback of the
maternal call occurred from a single, randomly chosen location. Chicks trained with
only a single spatial location showed some improvement in their learning of the mater-
nal call relative to chicks provided training involving frequent switching between loca-
tions within the 450–900 msec delay window (Figure 4). In Experiment 4, we explored
this hypothesis further, examining chick responsivity and orienting to the maternal call
during training, comparing trials involving playback from the same (Same trials) vs.
switch of location (Switch trials). We expected that chicks would perform better on
Same trials, generally, and would also have particular difficulty on Switch trials when
the stimulus fell 450–900 msec after their vocalizations. The first of these predictions
was well supported, as chicks showed superior orienting and localization of the mater-
nal call on Same relative to Switch trials (see Figure 5). Although chicks also tended
to be significantly closer to the correct location on Same compared to Switch trials,
our analysis of fluctuations in chick responsivity on the two types of trials (see Fig-
ure 6) suggests that spatial proximity cannot fully account for the observed superior
performance on Same trials. We argue that the most parsimonious explanation for our
results is one that draws on the available human literature on the differing temporal
dynamics of the “endogenous” and “exogenous” attentional systems as well as their
interaction (e.g., Corbetta & Shulman, 2002).
As our previous studies exploring the role of contingency in the formation of filial
preferences in quail chicks illustrate (e.g., Harshaw & Lickliter, 2007, 2011; Harshaw
et al., 2008), infants often respond vigorously to contingencies between their behaviors
and consequences (cf. Watson, 1972). Experimental paradigms capitalizing on this fact
are thus common in developmental work. For example, conjugate reinforcement para-
digms, which involve yoking a stimulus to an infant’s actions, have proven invaluable
as tools for the study of infant learning and memory (see Rovee-Collier & Cuevas,
2009), categorization (e.g., Bhatt, Wilk, Hill, & Rovee-Collier, 2004), and multisensory
integration (e.g., Kraebel, 2012). Infant-controlled habituation procedures, which
involve linking stimulation to infant gaze (Colombo & Horowitz, 1985), have similarly
provided invaluable research tools for developmentalists. More sophisticated gaze-con-
tingent paradigms, in which eye tracking is employed and stimuli are presented contin-
gent upon shifts in infant eye gaze, are also growing in popularity due to distinct
advantages over other methods (e.g., Barton & Bertenthal, 2013; Deligianni, Senju,
Gergely, & Csibra, 2011; Wang et al., 2012; Wass, Porayska-Pomsta, & Johnson,
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
TIMING-SPECIFIC DEFICIT IN AUDITORY LEARNING 23

2011). Although we are unaware of any human infant studies fully analogous to our
study—that is, employing contingent stimulus presentation, location switching, and
variable delays—there are many studies that employ contingent presentation and loca-
tion switching, often to control for side bias (e.g., Barton & Bertenthal, 2013). Our
findings demonstrate that the spatiotemporal parameters chosen in such paradigms
may potentially modify not only experimental outcomes, but the cognitive processes a
stimulus impinges upon in a task, given potential for interaction with infants’ develop-
ing attentional systems. One area where our results may have particular significance is
language, given the importance of social contingencies in language learning and the
frequent use of both contingency and location switching in studies of language devel-
opment (e.g., Gros-Louis, West, & King, in press; Lewkowicz, Minar, Tift, & Bran-
don, 2015; Roseberry et al., 2014). The hypothesis that results similar to those found
in quail chicks would be obtained in a similar study of human infants nonetheless
remains to be tested.
Although it is a truism in psychophysics and many areas of neuroscience that “tim-
ing is everything” when it comes to isolating the workings of a particular cognitive
process, thinking about the effects of delay on learning in operant paradigms is domi-
nated by the idea that the potency of reinforcers declines in a straightforward, hyper-
bolic fashion with increased delay (e.g., Mazur, 2001; Takahashi et al., 2007). Studies
showing violations of hyperbolic decline with delays of less than one-second have thus
tended to be ignored and/or stand as unexplained anomalies in the literature (e.g.,
Cerekwicki et al., 1969; Cerekwicki & Grant, 1967; Landauer, 1964; cf. Schneider,
1990). Our results indicate that one potential source for such violations is blind spots
or “blinks” caused by the spatiotemporal dynamics of attention in a given laboratory
task. Based on our findings, such timing-specific interference is likely to be maximal
under conditions where there is uncertainty regarding the location of an anticipated
event relative to the location of the subject or participant. Most studies of delay in
operant paradigms have nonetheless employed so few delays of less than a second that
the blink observed here would have been missed. The conditions created in our para-
digm—in which chicks can never be certain about the location of an impending stimu-
lus unless they learn a fairly complicated rule—are also seldom mirrored in standard
operant preparations, wherein animals (usually adults) are adept at learning where and
when their attention need be deployed (Gallistel & Gibbon, 2000).
We are unaware of any inherent ecological significance of the 450–900 msec “blink”
window documented here for bobwhite quail and further studies will be needed to
address this question. If avian attentional systems are prone to specific spatiotemporal
errors—confusions, blind spots or “blinks”—similar to those displayed by humans,
then either the costs of such errors must be low relative to benefits received or other
systems have evolved and/or emerge during ontogeny to defend against such costs.
Auditory signaling in the absence of visual contact with receivers is common in birds
and many other animals (e.g., Perez, Fernandez, Griffith, Vignal, & Soula, 2015). This
is the case for quail in the days after hatch, particularly as they move about and forage
while maintaining vocal contact with mother and broodmates (Stokes, 1967). It is also
true in mixed-species flocks of warblers and other birds during migration, which not
only forage together during stops but exhibit frequent, intermittent contact calling via
“chip notes” as they do so. Signalers also frequently displace themselves relative both
to previously signaled-from locations and receivers. If the “blink” documented here is
general, then this would have consequences for auditory interactions in such contexts,
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
24 HARSHAW &LICKLITER

particularly if the timing of consecutive signals is brief. From signalers being able to
cloak their spatial locations by manipulating the nervous systems of receivers (cf. Daw-
kins, 1982) to predators/aggressors exploiting the attentional limitations of signalers,
the current study suggests a number of questions concerning real-world social and
communicative dynamics that warrant exploration.
In summary, we describe and replicate a novel, timing-specific deficit or “blink” in
auditory learning in 1- to 3-day old quail hatchlings, occurring 450–900 msec after
chicks vocalize to hear playback of a maternal call in an operant, delay-of-reinforce-
ment paradigm. Our results indicate that this deficit was caused by the spatiotemporal
dynamics of chicks’ attentional systems, interacting with the high degree of uncertainty
created in our training paradigm, which involved frequent switches of spatial location
in an arena entirely devoid of landmarks. Day-old chicks trained with only a single
location thus showed improved learning within the 450–900 msec relatively to chicks
trained with switches of spatial location. Microanalysis of chick behavior during train-
ing in our original paradigm also revealed patterns of fluctuations during the two trial
types as well as trial-to-trial sequential dependencies supportive of an attentional
account of the timing-specific deficit in learning. Although it is unclear whether there
is anything but surface similarity between the timing-specific “blink” in learning
described here and the phenomenon of “attentional blink” in humans, argued by many
to reflect the inhibitory or “dark” side of attention (Nieuwenstein, Potter, & Theeuwes,
2009; Olivers & Meeter, 2008; Raymond, Shapiro, & Arnell, 1992; Schroeder & Laka-
tos, 2009), the possibility remains to be explored. Further studies will be needed, both
to ascertain the generality of this “blink” as well as to explore its relation to timing-
specific deficits in humans.
ACKNOWLEDGMENTS 4
Xxxxxxxxxx.
REFERENCES
Arbuckle, J. L., & Lattal, K. A. (1988). Changes in functional response units with briefly delayed reinforce-
ment. Journal of the Experimental Analysis of Behavior,49, 249–263.
Bahrick, L. E. (1995). Intermodal origins of self-perception. Advances in Psychology,112, 349–373.
Bahrick, L. E. (2013). Body perception: Intersensory origins of self and other perception in newborns. Cur-
rent Biology,23, R1039–R1041.
Bahrick, L. E., & Watson, J. S. (1985). Detection of intermodal proprioceptive–visual contingency as a
potential basis of self-perception in infancy. Developmental Psychology,21, 963–973.
Barton, A., & Bertenthal, B. (2013). Infants differentially anticipate the goals of ipsilateral and contralateral
reaches. Journal of Vision,13, 283. [Meeting abstract presented at VSS 2013].
Bays, P. M., & Husain, M. (2012). Active inhibition and memory promote exploration and search of natural
scenes. Journal of Vision,12,8.
Bhatt, R. S., Wilk, A., Hill, D., & Rovee-Collier, C. K. (2004). Correlated attributes and categorization in
the first half-year of life. Developmental Psychobiology,44, 103–115.
Blackburn, M., & El-Deredy, W. (2013). The future is risky: Discounting of delayed and uncertain outcomes.
Behavioural Processes,94,9–18.
Bloom, K. (1977). Patterning of infant vocal behavior. Journal of Experimental Child Psychology,23, 367–
377.
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
TIMING-SPECIFIC DEFICIT IN AUDITORY LEARNING 25

Bornstein, M. H., Putnick, D. L., Cote, L. R., Haynes, O. M., & Suwalsky, J. T. (2015). Mother-infant con-
tingent vocalizations in 11 countries. Psychological Science,26, 1272–1284.
Catania, A. C. (2006). Learning, 4th ed. New York: Sloan Publishing.
Celec, P. (2004). Analysis of rhythmic variance-ANORVA. A new simple method for detecting rhythms in
biological time series. Biological Research,37, 777–782.
Cerekwicki, L. E., & Grant, D. A. (1967). Delay of positive reinforcement in instrumental eyelid condition-
ing. Journal of Experimental Psychology,75, 360–364.
Cerekwicki, L. E., Kantowitz, B. H., & Grant, D. A. (1969). Replicability of an optimal delay of reinforce-
ment result in instrumental eyelid conditioning. Journal of Experimental Psychology,79, 189–190.
Collias, N. E. (1952). The development of social behavior in birds. The Auk,69, 127–159.
Colombo, J., & Horowitz, F. D. (1985). A parametric study of the infant control procedure. Infant Behavior
and Development,8, 117–121.
Corbetta, M., & Shulman, G. L. (2002). Control of goal-directed and stimulus-driven attention in the brain.
Nature Reviews Neuroscience,3, 201–215.
Dawkins, R. (1982). The extended phenotype. ???????: Oxford University Press. 5
Deligianni, F., Senju, A., Gergely, G., & Csibra, G. (2011). Automated gaze-contingent objects elicit orienta-
tion following in 8-month-old infants. Developmental Psychology,47, 1499–1503.
Evans, C. S. (1991). Of ducklings and Turing machines: Interactive playbacks enhance subsequent respon-
siveness to conspecific calls. Ethology,89, 125–134.
Filippetti, M. L., Johnson, M. H., Lloyd-Fox, S., Dragovic, D., & Farroni, T. (2013). Body perception in
newborns. Current Biology,23, 2413–2416.
Foush
ee, R. D., & Lickliter, R. (2002). Early visual experience affects postnatal auditory responsiveness in
bobwhite quail (Colinus virginianus). Journal of Comparative Psychology,116, 369–380.
Gallistel, C. R., & Gibbon, J. (2000). Time, rate, and conditioning. Psychological Review,107, 289.
Gallup, G. G., & Suarez, S. D. (1980). An ethological analysis of open-field behaviour in chickens. Animal
Behaviour,28, 368–378.
Gergely, G., & Watson, J. S. (1999). Early socio-emotional development: Contingency perception and the
social-biofeedback model. In P. Rochat (Ed.), Early social cognition: Understanding others in the first
months of life (pp. 101–135). Mahwah, NJ: Lawrence Erlbaum Associates.
Goldstein, M. H., King, A. P., & West, M. J. (2003). Social interaction shapes babbling: Testing parallels
between birdsong and speech. Proceedings of the National Academy of Sciences,100, 8030–8035.
Goldstein, M. H., & Schwade, J. A. (2008). Social feedback to infants’ babbling facilitates rapid phonological
learning. Psychological Science,19, 515–523.
Goldstein, M. H., & Schwade, J. A. (2010). From birds to words: Perception of structure in social interac-
tions guides vocal development and language learning. In M. S. Blumberg, J. H. Freeman, & S. R. Robin-
son (Eds.), The oxford handbook of developmental behavioral neuroscience (pp. 708–729). New York:
Oxford University Press.
Goldstein, M. H., Schwade, J. A., & Bornstein, M. H. (2009). The value of vocalizing: Five-month-old
infants associate their own noncry vocalizations with responses from caregivers. Child Development,80,
636–644.
Goldstein, M. H., Schwade, J., Briesch, J., & Syal, S. (2010). Learning while babbling: Prelinguistic object-
directed vocalizations indicate a readiness to learn. Infancy,15, 362–391.
Goldstein, M. H., Waterfall, H. R., Lotem, A., Halpern, J. Y., Schwade, J. A., Onnis, L., & Edelman, S.
(2010). General cognitive principles for learning structure in time and space. Trends in Cognitive Sciences,
14, 249–258.
Gottlieb, G. (1963). A naturalistic study of imprinting in wood ducklings (Aix sponsa). Journal of Compara-
tive and Physiological Psychology,56,86–91.
Gottlieb, G. (1971). Development of species identification in birds: An inquiry into the prenatal determinants of
perception. Chicago, IL: University of Chicago Press.
Gottlieb, G. (1987). Development of species identification in ducklings. XIII: A comparison of malleable and
critical periods of perceptual development. Developmental Psychobiology,20, 393–404.
Gottlieb, G. (1997). Synthesizing nature-nurture: Prenatal roots of instinctive behavior. ????????: Psychology
Press. 6
Gros-Louis, J., West, M. J., Goldstein, M. H., & King, A. P. (2006). Mothers provide differential feedback
to infants’ prelinguistic sounds. International Journal of Behavioral Development,30, 509–516.
Gros-Louis, J., West, M. J., & King, A. P. (In press). The influence of interactive context on prelinguistic
vocalizations and maternal responses. Language Learning and Development,???, ???–???. 7
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
26 HARSHAW &LICKLITER

Harshaw, C., & Lickliter, R. (2007). Interactive and vicarious acquisition of auditory preferences in Northern
bobwhite (Colinus virginianus) chicks. Journal of Comparative Psychology,121, 320–331.
Harshaw, C., & Lickliter, R. (2011). Biased embryos: Prenatal experience alters the postnatal malleability of
auditory preferences in bobwhite quail. Developmental Psychobiology,53, 291–302.
Harshaw, C., Tourgeman, I. P., & Lickliter, R. (2008). Stimulus contingency and the malleability of species-
typical auditory preferences in Northern bobwhite (Colinus virginianus) hatchlings. Developmental Psy-
chobiology,50, 460–472.
Heaton, M. B., Miller, D. B., & Goodwin, D. G. (1978). Species-specific auditory discrimination in bobwhite
quail neonates. Developmental Psychobiology,11,13–21.
Honeycutt, H., & Lickliter, R. (2001). Order-dependent timing of unimodal and multimodal stimulation
affects prenatal auditory learning in bobwhite quail embryos. Developmental Psychobiology,38,1–10.
Honeycutt, H., & Lickliter, R. (2002). Prenatal experience and postnatal perceptual preferences: Evidence for
attentional-bias in bobwhite quail embryos (Colinus virginianus). Journal of Comparative Psychology,116,
270–276.
Impekoven, M. (1973). Response-contingent prenatal experience of maternal calls in the Peking duck (Anas
platyrhynchos). Animal Behaviour,21, 164–168.
Jones, J. E. (1962). Contiguity and reinforcement in relation to CS-UCS intervals in classical aversive condi-
tioning. Psychological Review,69, 176–186.
Kirby, K. N., & Marakovi
c, N. N. (1996). Delay-discounting probabilistic rewards: Rates decrease as
amounts increase. Psychonomic Bulletin & Review,3, 100–104.
Klein, R. (2000). Inhibition of return. Trends in Cognitive Sciences,4, 138–147.
Kraebel, K. S. (2012). Redundant amodal properties facilitate operant learning in 3-month-old infants. Infant
Behavior and Development,35,12–21.
Landauer, T. K. (1964). Delay of an acquired reinforcer. Journal of Comparative and Physiological Psychol-
ogy,58, 374.
Lattal, K. A., & Ziegler, D. R. (1982). Briefly delayed reinforcement: An interresponse time analysis. Journal
of the Experimental Analysis of Behavior,37, 407–416.
Lewkowicz, D. J., Minar, N. J., Tift, A. H., & Brandon, M. (2015). Perception of the multisensory coherence
of fluent audiovisual speech in infancy: Its emergence and the role of experience. Journal of Experimental
Child Psychology,130, 147–162.
Lickliter, R., & Harshaw, C. (2010). Canalization and malleability reconsidered. In K. E. Hood, C. T. Hal-
pern, G. Greenberg, & R. M. Lerner (Eds.), Handbook of developmental science, behavior, and genetics (pp.
491–524). ????????: Wiley-Blackwell. 8
Lickliter, R., & Hellewell, T. B. (1992). Contextual determinants of auditory learning in bobwhite quail
embryos and hatchlings. Developmental Psychobiology,25,17–31.
Lorenz, K. Z. (1937). The companion in the bird’s world. The Auk,54, 245–273.
Mazur, J. E. (1987). An adjusting procedure for studying delayed reinforcement. In M. L. Commons, J. E.
Mazur, J. A. Nevin, & H. Rachlin (Eds.), Quantitative analyses of behavior: The effect of delay and of inter-
vening events on reinforcement value (Vol. 5, pp. ???–???). Hillsdale, NJ: Erlbaum. 9
Mazur, J. E. (2001). Hyperbolic value addition and general models of animal choice. Psychological Review,
108,96–112.
M€
uller, N. G., & Kleinschmidt, A. (2007). Temporal dynamics of the attentional spotlight: Neuronal corre-
lates of attentional capture and inhibition of return in early visual cortex. Journal of Cognitive Neuro-
science,19, 587–593.
Nieuwenstein, M. R., Potter, M. C., & Theeuwes, J. (2009). Unmasking the attentional blink. Journal of
Experimental Psychology: Human Perception and Performance,35, 159–169.
Odum, A. L., Baumann, A. A., & Rimington, D. D. (2006). Discounting of delayed hypothetical money and
food: Effects of amount. Behavioural Processes,73, 278–284.
Olivers, C. N. L., & Meeter, M. (2008). A boost and bounce theory of temporal attention. Psychological
Review,115, 836–863.
Perez, E. C., Fernandez, M. S. A., Griffith, S. C., Vignal, C., & Soula, H. A. (2015). Impact of visual contact
on vocal interaction dynamics of pair-bonded birds. Animal Behaviour,107, 125–137.
Posner, M. I., & Cohen, Y. (1984). Components of performance. In H. Bouma, & D. Bowhuis (Eds.), Atten-
tion and performance (Vol. X, pp. 531–556). Hillsdale, NJ: Psychology Press.
Raymond, J. E., Shapiro, K. L., & Arnell, K. M. (1992). Temporary suppression of visual processing in an
RSVP task: An attentional blink? Journal of Experimental Psychology: Human Perception and Performance,
18, 849–860.
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
TIMING-SPECIFIC DEFICIT IN AUDITORY LEARNING 27

Rescorla, R. A. (1988). Behavioral studies of Pavlovian conditioning. Annual Review of Neuroscience,11,
329–352.
Richards, R. W. (1981). A comparison of signaled and unsignaled delay of reinforcement. Journal of the
Experimental Analysis of Behavior,35, 145–152.
Robertson, S. S., Watamura, S. E., & Wilbourn, M. P. (2012). Attentional dynamics of infant visual forag-
ing. Proceedings of the National Academy of Sciences,109, 11460–11464.
Rochat, P., & Morgan, R. (1995). Spatial determinants in the perception of self-produced leg movements in
3-to 5-month-old infants. Developmental Psychology,31, 626–636.
Roseberry, S., Hirsh-Pasek, K., & Golinkoff, R. M. (2014). Skype me!Socially contingent interactions help
toddlers learn language. Child Development,85, 956–970.
Rovee-Collier, C. K., & Cuevas, K. (2009). The development of infant memory. In M. L. Courage, & N.
Cowan (Eds.), The development of memory in infancy and childhood (pp. 11–42). ???????: Taylor & Francis. 10
Rumpf, M., & Tzschentke, B. (2010). Perinatal acoustic communication in birds: Why do birds vocalize in
the egg? Open Ornithology Journal,3, 141–149.
Schneider, S. M. (1990). The role of contiguity in free-operant unsignaled delay of positive reinforcement: A
brief review. The Psychological Record,40, 239–257.
Schneider, S. M., & Harshaw, C. (2007). Whence malleability? Gottlieb’s ducklings, operant contingency,
and the social manifold. European Journal of Developmental Science,1, 233–237.
Schneider, S. M., & Lickliter, R. (2010). Operant generalization in quail neonates after intradimensional
training: Distinguishing positive and negative reinforcement. Behavioural Processes,83,1–7.
Schroeder, C. E., & Lakatos, P. (2009). Low-frequency neuronal oscillations as instruments of sensory selec-
tion. Trends in Neurosciences,32,9–18.
Schulz, D., & Huston, J. P. (2002). The sliding window correlation procedure for detecting hidden correla-
tions: Existence of behavioral subgroups illustrated with aged rats. Journal of Neuroscience Methods,121,
129–137.
Sizemore, O. J., & Lattal, K. A. (1978). Unsignalled delay of reinforcement in variable-interval schedules.
Journal of the Experimental Analysis of Behavior,30, 169–175.
Skinner, B. F. (1948). Superstition in the pigeon. Journal of Experimental Psychology,38, 168–172.
Stokes, A. W. (1967). Behavior of the bobwhite, Colinus virginianus. The Auk,84,1–33.
Stoumbos, J. A. (1990). Effects of altered prenatal auditory experience on postnatal auditory preferences in
bobwhite quail chicks. Unpublished master’s thesis, Blacksburg, Virginia: Virginia Polytechnic Institute.
Takahashi, T., Ikeda, K., & Hasegawa, T. (2007). A hyperbolic decay of subjective probability of obtaining
delayed rewards. Behavioral and Brain Functions,3, 52.
Timberlake, W. D., & Lucas, G. A. (1985). The basis of superstitious behavior: Chance contingency, stimulus
substitution, or appetitive behavior? Journal of the Experimental Analysis of Behavior,44, 279–299.
Tuculescu, R. A., & Griswold, J. G. (1983). Prehatching interactions in domestic chickens. Animal Behaviour,
31,1–10.
Wang, Q., Bolhuis, J., Rothkopf, C. A., Kolling, T., Knopf, M., & Triesch, J. (2012). Infants in control:
Rapid anticipation of action outcomes in a gaze-contingent paradigm. PLoS One,7, e30884.
Wass, S., Porayska-Pomsta, K., & Johnson, M. H. (2011). Training attentional control in infancy. Current
Biology,21, 1543–1547.
Watson, J. S. (1972). Smiling, cooing, and” the game”. Merrill-Palmer Quarterly of Behavior and Develop-
ment,18, 323–339.
Watson, J. S. (1979). Perception of contingency as a determinant of social responsiveness. In E. Thoman
(Ed.), The origins of social responsiveness (pp. 33–64). Hillsdale, NJ: Lawrence Erlbaum Associates.
Watson, J. S. (2001). Contingency perception and misperception in infancy: Some potential implications for
attachment. Bulletin of the Menninger Clinic,65, 296–320.
Weichselgartner, E., & Sperling, G. (1987). Dynamics of automatic and controlled visual attention. Science,
238, 778–780.
West, M. J., & King, A. P. (1988). Female visual displays affect the development of male song in the cow-
bird. Nature,334, 244–246.
Williams, B. A. (2001). The critical dimensions of the response-reinforcer contingency. Behavioural Processes,
54, 111–126.
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
28 HARSHAW &LICKLITER

Author Query Form
Journal: INFA
Article: 12139
Dear Author,
During the copy-editing of your paper, the following queries arose. Please respond to
these by marking up your proofs with the necessary changes/additions. Please write
your answers on the query sheet if there is insufficient space on the page proofs.
Please write clearly and follow the conventions shown on the attached corrections
sheet. If returning the proof by fax do not write too close to the paper’s edge. Please
remember that illegible mark-ups may delay publication.
Many thanks for your assistance.
Query
reference
Query Remarks
1 AUTHOR: Please confirm that given names (red) and
surnames/family names (green) have been identified
correctly.
2 AUTHOR: Please check that authors and their affili-
ations are correct.
3 AUTHOR: Please provide explanation for part figure
“f”in the caption of figure 3.
4 AUTHOR: Please provide “Acknowledgments”sec-
tion (if applicable).
5 AUTHOR: Please provide the publisher location for
reference Dawkins (1982).
6 AUTHOR: Please provide the publisher location for
reference Gottlieb (1997).
7 AUTHOR: Please provide the volume number, page
range for reference Gros-Louis et al. (In press).
8 AUTHOR: Please provide the publisher location for
reference Lickliter and Harshaw (2010).
9 AUTHOR: Please provide the page range for reference
Mazur (1987).
10 AUTHOR: Please provide the publisher location for
reference Rovee-Collier and Cuevas (2009).

11 AUTHOR: Figure 2 has been saved at a low resolution
of 259 dpi. Please resupply at 600 dpi. Check required
artwork specifications at http://authorservices.wiley.-
com/bauthor/illustration.asp
12 AUTHOR: Figure 3 has been saved at a low resolution
of 258 dpi. Please resupply at 600 dpi. Check required
artwork specifications at http://authorservices.wiley.-
com/bauthor/illustration.asp
13 AUTHOR: Figure 4 has been saved at a low resolution
of 167 dpi. Please resupply at 600 dpi. Check required
artwork specifications at http://authorservices.wiley.-
com/bauthor/illustration.asp
14 AUTHOR: Figure 5 has been saved at a low resolution
of 186 dpi. Please resupply at 600 dpi. Check required
artwork specifications at http://authorservices.wiley.-
com/bauthor/illustration.asp
15 AUTHOR: Figure 6 has been saved at a low resolution
of 192 dpi. Please resupply at 600 dpi. Check required
artwork specifications at http://authorservices.wiley.-
com/bauthor/illustration.asp
16 AUTHOR: Figure 7 has been saved at a low resolution
of 260 dpi. Please resupply at 600 dpi. Check required
artwork specifications at http://authorservices.wiley.-
com/bauthor/illustration.asp