Figure 1 - uploaded by Isabel Gauthier
Content may be subject to copyright.
Top panel: Targets for each of the three NOMTs (Greebles, Ziggerins, and Sheinbugs). Bottom panel: Illustration of NOMT test format (learning and test phase) with Ziggerins.
Source publication
In tests of object recognition, individual differences typically correlate modestly but nontrivially across familiar categories (e.g. cars, faces, shoes, birds, mushrooms). In theory, these correlations could reflect either global, non-specific mechanisms, such as general intelligence (IQ), or more specific mechanisms. Here, we introduce two separa...
Contexts in source publication
Context 1
... were created for three categories of novel objects (Greebles, Ziggerins, and Sheinbugs, see Figure 1). NOMTs were closely modeled after the CFMT (see Duchaine & Nakayama, 2006). ...
Context 2
... were closely modeled after the CFMT (see Duchaine & Nakayama, 2006). Each NOMT (see Figure 1) started with a learning phase, where a target object was shown in three views (3 seconds per view) followed by three test items where participants had to select which of three objects was the object they had just studied. This was repeated for each of six target objects. ...
Citations
... Participants completed an online cognitive assessment measuring processing speed, visual short-term memory, attention, cognitive control, working memory, perception, general intelligence, and verbal reasoning. Six tests from the TestMyBrain digital platform 32,33 were used: (1) TestMyBrain Digit Symbol Matching; [33][34][35] (2) TestMyBrain Gradual Onset Continuous Performance Test; 33,36,37 (3) TestMyBrain Verbal Paired Associates; 33,38 (4) TestMyBrain Visual Paired Associates; 33 (5) TestMyBrain Multiple Object Tracking; 37 (6) TestMyBrain Vocabulary 35,39 (eTable 2 describes the tests and the cognitive domains covered by each test). Both Verbal and Visual Paired tests are delayed recall tests (2-3 min for verbal pairs, 4-5 min for visual pairs) that have been used in prior studies of visual and verbal long-term memory 33,40 . ...
Childhood cognitively stimulating activities have been associated with higher cognitive function in late life. Whether activities in early or late childhood are more salient, and whether activities are associated with specific cognitive domains is unknown. Participants retrospectively reported cognitively stimulating activities at ages 6, 12, and 18 years. 4,198 participants were aged 55 to 77 years at cognitive testing. Six tasks measured overall cognitive function, processing speed, visual short-term memory, attention, cognitive control, episodic memory, working memory, perception, vocabulary, and verbal reasoning. Cognitively stimulating activities across childhood were associated with higher cognitive scores (highest versus lowest quartile, beta = 0.18 SD, 95% CI = 0.12, 0.23). In models adjusted for activities at each age, only age 18 activities were associated with overall cognition. The association of activities with cognitive function was strongly positive at the lowest levels of activities, with little association at middle and high levels of activities. A test of crystalized intelligence was most strongly associated with activities; tests assessing processing speed, visual short-term memory, visual working memory, and sustained attention were least associated. If the associations we found are causal, increasing cognitively stimulating activities in the late teen years among those with very few activities may benefit late life cognitive health.
Supplementary Information
The online version contains supplementary material available at 10.1038/s41598-024-79083-x.
... One important highlevel visual ability is object recognition. This is the ability to discriminate between visually similar objects, which contributes to performance on a variety of tasks including matching, memory judgments or even judgments about summary statistics for groups of objects (Richler et al., 2017(Richler et al., , 2019Sunday et al., 2021). The novel object memory tests (NOMTs, Richler et al., 2017) are simple tests that avoid the influence of category-specific experiences on visual tasks (e.g., birding classes or attending car shows). ...
... The novel object memory tests (NOMTs, Richler et al., 2017) are simple tests that avoid the influence of category-specific experiences on visual tasks (e.g., birding classes or attending car shows). According to one estimate, almost none of the variance shared between different NOMTs is explained by IQ (Richler et al., 2017). NOMTs have been used as indicators of a domain-general ability that contributes unique variance (beyond IQ) to the prediction of complex visual tasks like reading musical notation (Chang & Gauthier, 2021) or learning to make a variety of medical decisions that rely on visual information (Holmes et al., 2020;Smithson et al., 2022;Sunday et al., 2018aSunday et al., , 2018b. ...
... There is no a priori reason to expect that object recognition measures should be biased according to gender or ethnicity, and these tests cannot be easily adapted in different languages, given they include minimal verbal instructions. With the exception of data on age and gender (Richler et al., 2017), there is little known about potential sources of bias, and the role of language in visual tests has not been studied before. ...
Measurement of object recognition (OR) ability could predict learning and success in real-world settings, and there is hope that it may reduce bias often observed in cognitive tests. Although the measurement of visual OR is not expected to be influenced by the language of participants or the language of instructions, these assumptions remain largely untested. Here, we address the challenges of measuring OR abilities across linguistically diverse populations. In Study 1, we find that English–Spanish bilinguals, when randomly assigned to the English or Spanish version of the novel object memory test (NOMT), exhibit a highly similar overall performance. Study 2 extends this by assessing psychometric equivalence using an approach grounded in item response theory (IRT). We examined whether groups fluent in English or Spanish differed in (a) latent OR ability as assessed by a three-parameter logistic IRT model, and (2) the mapping of observed item responses on the latent OR construct, as assessed by differential item functioning (DIF) analyses. Spanish speakers performed better than English speakers, a difference we suggest is due to motivational differences between groups of vastly different size on the Prolific platform. That we found no substantial DIF between the groups tested in English or Spanish on the NOMT indicates measurement invariance. The feasibility of increasing diversity by combining groups tested in different languages remains unexplored. Adopting this approach could enable visual scientists to enhance diversity, equity, and inclusion in their research, and potentially in the broader application of their work in society.
... A significant proportion of face discrimination ability in our foraging task could be explained by demographics and performance on control trials, where older and less educated people in general were poorer at face discrimination than those younger and more educated, and those who did well on foraging for targets defined by simple features (black targets among white distractors or vice versa) tended to also be good at foraging for faces. These results may at least in part be due to various nonspecific factors that correlate with one or more of these variables, such as motor control, attention, motivation, foraging organization (Ólafsdóttir et al., 2021), intelligence (Richler et al., 2017;2019), and familiarity with computer-based tasks. Face discrimination ability was however additionally surprisingly well predicted by visual discrimination ability for miscellaneous other objects (compare to Richler et al., 2019), in accordance with individual differences in a general object perception factor o . ...
What are the diagnostic dimensions on which objects differ visually? We constructed a two-dimensional object space based on such attributes captured by a deep convolutional neural network. These attributes can be approximated as stubby/spiky and animate-/inanimate-looking. If object space contributes to human visual cognition, this should have a measurable effect on object discrimination abilities. We administered an object foraging task to a large, diverse sample (N = 511). We focused on the stubby animate-looking “face quadrant” of object space given known variations in face discrimination abilities. Stimuli were picked out of tens of thousands of images to either match or not match with the coordinates of faces in object space. Results show that individual differences in face perception can to a large part be explained by variability in general object perception abilities (o-factor). However, additional variability in face processing can be attributed to visual similarity with faces as captured by dimensions of object space; people who struggle with telling apart faces also have difficulties with discriminating other objects with the same object space attributes. This study is consistent with a contribution of object space to human visual cognition.
... PS is a construct related to general intelligence (Ackerman et al., 2002;Redick et al., 2013;Salthouse & Babcock, 1991) and is well-positioned as a potential source of shared variance between EP and OR. In prior studies (Richler et al., 2017(Richler et al., , 2019, measures of general intelligence like Raven's progressive matrices (Raven, 2000) have either not correlated strongly with general OR or were not able to account for all of the stable variance in its measurement (Smithson et al., 2024). ...
... Following prior work (Chang & Gauthier, 2022;Gauthier & Fiestan, 2023), the tests use novel objects and different task demands, and their aggregate provides an estimate of the latent factor o, as measured with CFA with a larger set of measures (Smithson et al., 2024). One measure was the novel object memory test-greeble (Richler et al., 2017; Figure 5A). In this test, participants completed three study-test blocks where they first studied six target symmetric greebles for 20 s in a study phase and then on each trial, chose which of three greebles was one of the six targets. ...
People can report summary statistics for various features about a group of objects. One theory is that different abilities support ensemble judgments about low-level features like color versus high-level features like identity. Existing research mostly evaluates such claims based on evidence of correlations within and between feature domains. However, correlations between two identical tasks that only differ in the type of feature that is used can be inflated by method variance. Another concern is that conclusions about high-level features are mostly based on faces. We used latent variable methods on data from 237 participants to investigate the abilities supporting low-level and high-level feature ensemble judgments. Ensemble judgment was measured with six distinct tests, each requiring judgments for a distinct low-level or high-level feature, using different task requirements. We also controlled for other general visual abilities when examining how low-level and high-level ensemble abilities relate to each other. Confirmatory factor analyses showed a perfect correlation between the two factors, suggesting a single ability. There was a unique relationship between these two factors beyond the influence of object recognition and perceptual speed. Additional results from 117 of the same participants also ruled out the role of working memory. This study provides strong evidence of a general ensemble judgment ability across a wide range of features at the latent level and characterizes its relationship to other visual abilities.
... To identify a truly domaingeneral object recognition ability, it is necessary to account for variance caused by (1) experience with specific object categories, (2) the kinds of features diagnostic for objects in specific categories, and (3) the specific demands of the tasks used to measure performance. To eliminate experience-related variance, novel categories of objects such as Greebles (Gauthier & Tarr, 1997) can be used, as is done in the Novel Object Memory Test (Richler et al., 2017). Alternatively, latent variable modeling or the aggregation of scores across tests can combine performance for multiple familiar object categories, minimizing the influence of variation in experience and any other property of a specific category (e.g., whether texture is diagnostic, or whether the objects are symmetrical). ...
... Although a two-sided Fisher's z-test (.82-.75; z = 1.878, p = .06; 95% CI [−0.03, 0.15]) was not significant, the difference is consistent with previous evidence that novel object recognition tests correlate more highly with each other than do familiar object recognition tests (Richler et al., 2017), potentially due to a greater role for domain-general ability in tests of novel objects. Given these results, we can be confident that for both theoretical and empirical reasons, using novel objects to estimate o may be better than using familiar objects, or no different, but are unlikely to provide worse estimates. ...
... In prior work using the aggregate approach to estimate o, novel object tasks have been used to avoid any contribution from experience with familiar objects. A large and varied set of familiar categories may converge on an unbiased estimate of domain-general ability (Richler et al., 2017), but measurement with novel objects may provide a more direct way to avoid this potential source of bias. ...
Measurement of domain-general object recognition ability (o) requires minimization of domain-specific variance. One approach is to model o as a latent variable explaining performance on a battery of tests which differ in task demands and stimuli; however, time and sample requirements may be prohibitive. Alternatively, an aggregate measure of o can be obtained by averaging z-scores across tests. Using data from Sunday et al., Journal of Experimental Psychology: General, 151, 676–694, (2022), we demonstrated that aggregate scores from just two such object recognition tests provide a good approximation (r = .79) of factor scores calculated from a model using a much larger set of tests. Some test combinations produced correlations of up to r = .87 with factor scores. We then revised these tests to reduce testing time, and developed an odd one out task, using a unique object category on nearly every trial, to increase task and stimuli diversity. To validate our measures, 163 participants completed the object recognition tests on two occasions, one month apart. Providing the first evidence that o is stable over time, our short aggregate o measure demonstrated good test–retest reliability (r = .77). The stability of o could not be completely accounted for by intelligence, perceptual speed, and early visual ability. Structural equation modeling suggested that our tests load significantly onto the same latent variable, and revealed that as a latent variable, o is highly stable (r = .93). Aggregation is an efficient method for estimating o, allowing investigation of individual differences in object recognition ability to be more accessible in future studies.
... Scientists study perception because it is important to human health, productivity, and quality of life. Recent progress has been spurred by the creation of tests designed to reliably measure high-level visual abilities in the normal population (Duchaine & Nakayama, 2006;McGugin et al., 2012;Richler et al., 2017). When many reliable tasks are available for a given construct, they can be used together as indicators of this construct in latent variable modelling (Meyer et al., 2021;Russell et al., 1998;Tomarken & Waller, 2005). ...
... In the visual Novel Object Memory Test, participants were tasked with remembering six target objects to later select them against arrays of foils regardless of viewpoint or visual noise (Fig. 1c). This test used Sheinbugs stimuli (Richler et al., 2017), which are 3D-rendered stimuli with a head, a front protrusion, and two arms; the item identity was defined by the unique combination of each component. This test began with a study phase where six Sheinbugs were presented simultaneously to study for as long as desired. ...
... In prior work with a similar sample, we found Bayesian evidence supporting the absence of a correlation between general intelligence and both the hMatch-Spaceships and the hNOMT-buttons (Chow et al., 2022a). Further, prior work with our visual tasks found that general intelligence did not account for the shared variance between them (Richler et al., 2017(Richler et al., , 2019 nor did including general intelligence in a model reduce how well o v predicted learning in a medical diagnostic task (Sunday et al., 2018). There is, however, some evidence from aging research that cross-modal hapticvisual performance can be related to a decline in cognitive abilities (Kalisch et al., 2012). ...
People vary in their ability to recognize objects visually. Individual differences for matching and recognizing objects visually is supported by a domain-general ability capturing common variance across different tasks (e.g., Richler et al., Psychological Review, 126, 226–251, 2019). Behavioral (e.g., Cooke et al., Neuropsychologia, 45, 484–495, 2007) and neural evidence (e.g., Amedi, Cerebral Cortex, 12, 1202–1212, 2002) suggest overlapping mechanisms in the processing of visual and haptic information in the service of object recognition, but it is unclear whether such group-average results generalize to individual differences. Psychometrically validated measures are required, which have been lacking in the haptic modality. We investigate whether object recognition ability is specific to vision or extends to haptics using psychometric measures we have developed. We use multiple visual and haptic tests with different objects and different formats to measure domain-general visual and haptic abilities and to test for relations across them. We measured object recognition abilities using two visual tests and four haptic tests (two each for two kinds of haptic exploration) in 97 participants. Partial correlation and confirmatory factor analyses converge to support the existence of a domain-general haptic object recognition ability that is moderately correlated with domain-general visual object recognition ability. Visual and haptic abilities share about 25% of their variance, supporting the existence of a multisensory domain-general ability while leaving a substantial amount of residual variance for modality-specific abilities. These results extend our understanding of the structure of object recognition abilities; while there are mechanisms that may generalize across categories, tasks, and modalities, there are still other mechanisms that are distinct between modalities.
... We will refer to this ability as o (Richler et al., 2019), but note that a general factor in the visual domain has also been referred to as VG (Hendel, Starrfelt, & Gerlach, 2019). A visual o-factor may explain a large part of individual differences in object perception, is at least partially modality-specific (Chow, Palmeri, & Gauthier, 2024) but see (Chow, Palmeri, Pluck, & Gauthier, 2023), and appears to be separable from low-level visual abilities as well as other cognitive abilities such as general intelligence (Chow et al., 2023;Richler et al., 2019;Richler, Wilmer, & Gauthier, 2017). Individual differences in o are related to neural measures of shape selectivity, including in suggested categoryselective regions of the ventral visual pathway (McGugin, Sunday, & Gauthier, 2023). ...
... A significant proportion of face discrimination ability in our foraging task could be explained by demographics and performance on control trials, where older and less educated people in general were poorer at face discrimination than those younger and more educated, and those who did well on foraging for targets defined by simple features (black targets among white distractors or vice versa) tended to also be good at foraging for faces. These results may at least in part be due to various nonspecific factors that correlate with one or more of these variables, such as motor control, attention, motivation, foraging organization (Ólafsdóttir et al., 2021), intelligence (Richler et al., 2017;2019), and familiarity with computer-based tasks. Face discrimination ability was however additionally surprisingly well predicted by visual discrimination ability for miscellaneous other objects (compare to Richler et al., 2019), in accordance with individual differences in a general object perception factor o (Hendel et al., 2019;Richler et al., 2019). ...
What are the diagnostic dimensions on which objects truly differ visually? We constructed a two-dimensional object space based on such attributes captured by a deep convolutional neural network. These attributes can be approximated as stubby/spiky and animate-/inanimate-looking. If object space underlies human visual cognition, this should have a measurable effect on object discrimination abilities. We administered an object foraging task to a large, diverse sample (N=511). We focused on the stubby animate-looking “face quadrant” of object space given known variations in face discrimination abilities. Stimuli were picked out of tens of thousands of images to either match or not match with the coordinates of faces in object space. People who struggled with telling apart faces also had difficulties with discriminating other objects with the same object space attributes. This study provides the first behavioral evidence for the existence of an object space in human visual cognition.
Public Significance Statement
The study emphasizes individual differences in visual cognition, a relatively neglected field of research. Unlike differences in other cognitive traits (e.g., Big Five personality traits, g-factor of general intelligence), we have limited knowledge on how people differ in their object processing capacity, and whether such abilities are fractionated or unitary. In this study, we ask whether visual object perception abilities are organized around an object space as evidenced by individual differences in behavior.
... This protocol includes a training block as well as pre-and post-training retinal image recognition tasks. Additionally, we examine any potential associations between the observers' ability to learn the retinal biomarkers and their domain-general visual object recognition ability, which we assess separately using a novel object memory task [35]. ...
... We examined domain-general object recognition ability and sex-recognition performance based on retinal fundus images in a group of experts and a group of naive observers. Assessment of object recognition ability was based on accuracy on the Novel Object Memory test (NOMT) using the Ziggerins novel object category [35]. Sex recognition task was completed twice, once before, and again after a training block. ...
... The methodological details of the NOMT test is described in Richler et al. [35], which we will summarize here. The NOMT is a 72-trial 3-AFC protocol which consists of a 18-trial learning phase and a 54-trial test phase. ...
We present a structured approach to combine explainability of artificial intelligence (AI) with the scientific method for scientific discovery. We demonstrate the utility of this approach in a proof-of-concept study where we uncover biomarkers from a convolutional neural network (CNN) model trained to classify patient sex in retinal images. This is a trait that is not currently recognized by diagnosticians in retinal images, yet, one successfully classified by CNNs. Our methodology consists of four phases: In Phase 1, CNN development, we train a visual geometry group (VGG) model to recognize patient sex in retinal images. In Phase 2, Inspiration, we review visualizations obtained from post hoc interpretability tools to make observations, and articulate exploratory hypotheses. Here, we listed 14 hypotheses retinal sex differences. In Phase 3, Exploration, we test all exploratory hypotheses on an independent dataset. Out of 14 exploratory hypotheses, nine revealed significant differences. In Phase 4, Verification, we re-tested the nine flagged hypotheses on a new dataset. Five were verified, revealing (i) significantly greater length, (ii) more nodes, and (iii) more branches of retinal vasculature, (iv) greater retinal area covered by the vessels in the superior temporal quadrant, and (v) darker peripapillary region in male eyes. Finally, we trained a group of ophthalmologists (N=26) to recognize the novel retinal features for sex classification. While their pretraining performance was not different from chance level or the performance of a nonexpert group (N=31), after training, their performance increased significantly (p<0.001, d=2.63). These findings showcase the potential for retinal biomarker discovery through CNN applications, with the added utility of empowering medical practitioners with new diagnostic capabilities to enhance their clinical toolkit.
... More information on this test can be found in the supplementary materials. An improved version of that test, which involved the sounds of keyboard switches, will be used in Experiment 2. Richler et al., (2017). The Greebles (Gauthier & Tarr, 1997) are a set of 3D novel objects defined by their body shape and their protrusions (with parts unique for each object and in this set, positioned in an asymmetrical configuration). ...
A general object recognition ability predicts performance across a variety of high-level visual tests, categories, and performance in haptic recognition. Does this ability extend to auditory recognition? Vision and haptics tap into similar representations of shape and texture. In contrast, features of auditory perception like pitch, timbre, or loudness do not readily translate into shape percepts related to edges, surfaces, or spatial arrangement of parts. We find that an auditory object recognition ability correlates highly with a visual object recognition ability after controlling for general intelligence, perceptual speed, low-level visual ability, and memory ability. Auditory object recognition was a stronger predictor of visual object recognition than all control measures across two experiments, even though those control variables were also tested visually. These results point towards a single high-level ability used in both vision and audition. Much work highlights how the integration of visual and auditory information is important in specific domains (e.g., speech, music), with evidence for some overlap of visual and auditory neural representations. Our results are the first to reveal a domain-general ability, o, that predicts object recognition performance in both visual and auditory tests. Because o is domain-general, it reveals mechanisms that apply across a wide range of situations, independent of experience and knowledge. As o is distinct from general intelligence, it is well positioned to potentially add predictive validity when explaining individual differences in a variety of tasks, above and beyond measures of common cognitive abilities like general intelligence and working memory.
... The second visual test was the Novel Objects Memory Test (NOMT, Richler et al., 2017) with one specific category of novel objects that share a common arrangement of parts (symmetrical Greebles, details of the task in Richler et al., 2017). Participants studied six different Greebles shown together on the screen for 20 sec, 3 times during the test, which included 48 trials 3-alternative forced choice trials with 3 Greebles shown on the screen until one was selected as 1 of the 6 studied targets. ...
... The second visual test was the Novel Objects Memory Test (NOMT, Richler et al., 2017) with one specific category of novel objects that share a common arrangement of parts (symmetrical Greebles, details of the task in Richler et al., 2017). Participants studied six different Greebles shown together on the screen for 20 sec, 3 times during the test, which included 48 trials 3-alternative forced choice trials with 3 Greebles shown on the screen until one was selected as 1 of the 6 studied targets. ...
Color is considered important in food perception, but its role in food-specific visual mechanisms is unclear. We explore this question in North American adults. We build on work revealing contributions from domain-general and domain-specific abilities in food recognition and a negative correlation between the domain-specific component and food neophobia (FN, aversion to novel food). In Study 1, participants performed two food-recognition tests, one in color and one in grayscale. Removing color reduced performance, but food recognition was predicted by domain-general and -specific abilities, and FN negatively correlated with food recognition. In Study 2, we removed color from both food tests. Food recognition was still predicted by domain-general and food-specific abilities, but with a relation between food-specific ability and FN. In Study 3, color-blind men reported lower FN than men with normal color perception. These results suggest two separate food-specific recognition mechanisms, only one of which is dependent on color.
