Humans show high proficiency in invariant object recognition, the ability to recognize the same objects from different viewpoints or in different scenes. This ability is supported by the ventral visual stream, the so-called what stream (Logothetis and Sheinberg, 1996). A question that is repeatedly addressed in vision studies is whether and how we can model this stream by means of animal models or computational models to further examine and quantify the representations along the ventral visual stream. Computationally, researchers have recently modelled this stream by using convolutional deep neural networks (CNNs), as, for example, done by Avberšek et al., 2021, Cadieu et al., 2014, Duyck et al., 2021, Güçlü and van Gerven, 2015, Kalfas et al., 2018, Kar et al., 2019, Kubilius et al., 2016, Pospisil et al., 2018, and Vinken and Op de Beeck, 2021. Lately, the rodent model has become an important animal model in vision studies, motivated by the applicability of molecular and genetic tools rather than by the visual capabilities of rodents. Past studies have examined behavioural (Alemi-Neissi et al., 2013; De Keyser et al., 2015; Djurdjevic et al., 2018; Schnell et al., 2019; Tafazoli et al., 2012; Vermaercke and Op de Beeck, 2012; Vinken et al., 2014; Zoccolan, 2015; for a review, see Zoccolan, 2015) as well as neural (Matteucci et al., 2019; Tafazoli et al., 2017; Vermaercke et al., 2014; Vinken et al., 2016) data of rodents (rats and mice) performing in visual pattern recognition tasks. The behavioural findings suggested that rats are capable of learning complex visual discrimination tasks. Here we plan to integrate computational and animal modelling approaches by using data about information processing in artificial neural networks when designing the animal experiments.

One aspect that almost all rodent studies have in common is that the exact task and stimuli are chosen based on what we know from human and monkey studies. Earlier research showed that the intuition of researchers about the complexity of visual tasks can be misleading (Vinken and Op de Beeck, 2021). Through computational CNN modelling of the tasks from previous studies, they showed that behavioural strategies that seem complex at first hand might be best modelled through relatively early levels of processing in CNNs. They recommended that future studies could obtain more direct information about the complexity of visual tasks and behavioural strategies by incorporating neural network models in the design phase of the experiment. One way of implementing this is to train rodents in a challenging and multidimensional visual task and use CNNs to select stimulus examples targeting strategies with different levels of complexity.

In this study, we implemented this approach and created a large stimulus set that can be used for a variety of visual experiments. We decided to create the stimuli in a way that they are adaptable to different types of tasks, such as a ‘simple’ discrimination task or non-linear tasks (e.g. Bossens and Op de Beeck, 2016). We then took a subset of these stimuli and performed a visual discrimination experiment in rats (see Figure 1 for the design). The task itself was defined in a stimulus space with two dimensions, here referred to as concavity and alignment. The stimuli consisted of a base shape that varied in concavity, with three spheres attached to it that were either horizontally aligned or misaligned. The task was then further complicated by transforming the stimuli along several dimensions that preserve the identity of the object. We started by training the animals in a base stimulus pair, with the target being the concave object with horizontally aligned spheres. Once the animals were trained in this base stimulus pair, we used the identity-preserving transformations to test for generalization. After a number of transformation phases, we selected a final stimulus set by choosing a combination of transformations based on the outcomes of a trained CNN. Using the neural network as a (basic) model for the different stages of ventral visual stream processing, we chose stimulus pairs that require either higher or lower levels of processing and thus allow us to maximally differentiate between the task strategies used by the animals. As a final part of this study, we performed an online human experiment with the same stimuli and design as the experiment for the rats, providing us with a rich three-way comparison of rat behavioural data with human behavioural data and with CNN data.

Figure 1 with 3 supplements see all

Download asset Open asset

In this study, we trained and tested 11 rats and 45 humans on a complex two-dimensional discrimination task (see Figure 1 for the design of the rat study and Figure 1—figure supplement 1 for the design of the human study). Rats and humans were first trained in a base pair. Next, we tested their ability to generalize across several image transformations. In the last two protocols of the design, we used a computational approach to select stimuli that require different visual strategies.

Animal study

Training

We first checked the variation in performance across phases and stimulus pairs during training. In the first training phase, animals were trained in the base stimulus pair, which were the maximally different target and distractor in a concavity × alignment stimulus space where each dimension was varied with four values (4 × 4 space). This training was successful for all 12 animals and lasted on average for 8.62 sessions (SD = 1.61). Animals were trained until they reached 80% performance for two consecutive sessions.

Once the animals were successfully trained, we examined whether they use both dimensions (concavity and alignment) by presenting them with two additional stimuli pairs where the target and distractor differ in only one dimension (see Figure 2, dimension learning). Performance on the old pair was similar to training performance (85.83%). The animals performed well with the stimuli that differ only along the concavity dimension (78.79%), although it was significantly lower than the performance on the base pair (paired t-test on rat performance, t(11) = 3.77, p=0.003). Performance dropped to 67.83% for the alignment-only pair, yet also significantly higher than chance level (one-sample t-test, p<0.0001). Overall, the dimension learning protocol provides evidence that the animals have picked up each of the two dimensions. This finding already excludes trivial explanations in terms of simple visual dimensions. For example, while concavity is correlated with horizontal size (distractor wider) and with overall brightness (distractor brighter, thus the opposite relevance as in the shaping phase), these simple dimensions cannot explain above-chance performance on the alignment dimension.

Figure 2

Download asset Open asset

The third training protocol consisted of a number of small transformations, as visualized in Figure 1 (transformations). Rats learned these transformations very well, with an average performance of 83.05% (see Figure 2). The pairwise percentage matrix in Figure 2 shows that the distractor with the size transformation (rightmost column in the matrix) affected the rat performance the most.

The variation in performance across targets and distractors can be due to a variety of factors. This can include simple dimensions such as brightness. In the base pair, the distractor is brighter than the target. While this is the opposite from the shaping task of detecting a shape versus a black screen, visual inspection of Figure 2 suggests that the animals perform poorer on trials in which the distractor display is not so much brighter (e.g. when it is small). To quantify this effect of brightness, we calculated the correlation between the performances in the matrix and the difference in pixel values (and thus brightness) of the stimulus pairs. This resulted in a (Pearson) correlation of –0.59 (p<0.01), suggesting that there is indeed an effect of brightness. Yet, brightness is at best a partial explanation because all percentages in the matrix are above chance, with the lowest percentage in the matrix being 68.83%, even though in some pairs the difference in pixel values is abolished or even opposite from the base pair.

Overall, the findings from the training phase and the above-chance performance on a variety of dimensions and transformations suggest that the rats have learned a pattern classification task with a level of complexity that might be competitive with other tasks in the rodent literature.

The six protocols that test generalization to various transformations with new, untrained images are associated with performances lower than 80% (binomial test, see Supplementary file 1a [lower table] for detailed table with results), but significantly higher than chance level (see Supplementary file 1a [lower table]). The pairwise percentage matrices of the animals in Figure 3 provide a more detailed view of what is happening in every test, and Figure 3—figure supplement 1 shows the individual accuracy for each animal. The distractor has a higher impact on performance than the target in some tests. Supplementary file 1b shows the marginal means and standard deviation for each target and distractor for these two test protocols. From these means it is clear that there is a higher variation in the performance between distractors in rotation X (52–65%) and rotation Z (56–73%) than between targets (55–60% resp. 60–66%). The same happens in the size test protocol.

Figure 3 with 1 supplement see all

Download asset Open asset

After these first six test protocols, the animals were presented with a schedule where all three rotations are combined (see Figure 1). On the new stimuli, the animals performed 58.56%, which is rather low, but still significantly different from chance level (binomial test on pooled performance of all animals: p<0.0001; 95% CI [0.57;0.60]).

Testing computational levels of complexity

For the final two test protocols, we used a CNN to find image pairs that would contrast strategies based upon a different stage in visual processing, with either early layers having lower performance than high layers (zero vs. high), or early layers having better performance than high layers (high vs. zero). Rat performance was particularly low for zero vs. high (56.47%), yet still significantly different from chance level when averaged across all stimulus pairs (binomial test on pooled performance of all animals; p<0.0001; 95% CI [0.55;0.58]). In contrast, rats were able to solve the high vs. zero pairs not only better than chance (average: 64.84%; binomial test on pooled performance of all animals; p<0.0001; 95% CI [0.63;0.66]), but also significantly better than zero vs. high (paired t-test on rat performance, t(10) = –4.49, p=0.0012). This suggests that rats align with lower levels of processing when we purposely select image pairs that are optimized to contrast different levels of the visual processing hierarchy.

Next we checked how much individual CNN layers can predict the variation in behavioural performance across image pairs when we take all test protocols together. We calculated the correlation of the generalization across image pairs between the CNN classifier (summarized in Figure 8) and the rat performance of all nine test protocols. This correlation includes a total of 287 image pairs, that is, all image pairs of all nine test protocols together. We did this by concatenating all performances of the animals into one array and all classification scores of the network into another array, and calculating the correlation between these two arrays to retrieve a correlation for each network layer. The results are displayed in Figure 4. Overall, we see quite low correlations, but several convolutional layers nevertheless show a significant positive correlation (permutation test) with the behavioural pattern of performance at the image pair level.

Figure 4

Download asset Open asset

Even though some of the correlations are significant, they are low. This could indicate that no CNN layer is able to capture what rats do. Alternatively, it could be caused by a very low reliability of the behavioural data. To test the reliability of the variations in behavioural performance between stimulus pairs in all nine test protocols, we calculated the split-half reliability, as previously done in Schnell et al., 2023, resulting in a correlation of 0.40. By applying the Spearman–Brown correction, we obtain a full-set reliability correlation of 0.58. This correlation is much higher than the correlations with individual CNN layers.

It is possible that rat performance would be based upon multiple levels of processing, in which case we would need a combination of layers in order to explain the variation in performance across stimulus pairs. Given the low correlation between neighbouring layers (Supplementary file 1c), a multiple linear regression was calculated with the classification scores of the 13 layers as 13 regressors, and the rat performances as response vector. The results of this regression indicate a significant effect of the classification scores (F(287,273) = 2.22, p=0.00907, R² = 0.10). Further investigating the 13 predictors showed that the later convolutional layers 8–10 of the network were significant predictors in the regression model (see Supplementary file 1d for results of the regression model). The R² = 10 of the full model would correspond to a correlation of around 0.32. This is better than the correlation of single layers, but still clearly smaller than the reliability of the rat data of 0.58. In conclusion, the CNN model provides a partial explanation of how the performance of rats varies across image pairs.

Given the relevance of convolutional layers, we can expect that relatively basic visual features might partially explain the behavioural strategy of rats. This includes dimensions such as brightness and pixel-based similarity. To get a first indication of the relevance of these features, we calculated the correlation across image pairs between rat performance and brightness and pixel similarity. Here we found a correlation of 0.34 for pixel similarity and 0.39 for brightness, suggesting that these two visual features partially explain our results when compared to the full-set reliability of rat performance (0.58).

Human study

A final part of this study was to include an online human study that follows the same design as the animal part. Figure 5 shows the average performance of humans (dark blue) versus rats (light blue) for all nine test protocols, as well as their performance on the old stimuli that were added in (or during) the testing protocols as quality control. Overall, humans performed better on all tests protocols than rats, with an average performance over all tests of 94.34% (humans) and 62.29% (rats). There was already a difference in terms of training performance (humans 92.86% vs. rats 77.84%), but the difference on the test protocols is larger. We subtracted the training performance of humans or rats from the testing performance of humans or rats, respectively, and even with this normalization for training performance there is still a significantly higher test performance in humans compared to rats (t(16) = –6.47, p<0.0001). Thus, not surprisingly, the degree of invariance in this object classification task is higher for humans compared to rat.

Figure 5

Download asset Open asset

The variation in performance across test protocols and across image pairs can give an indication of the strategies that each species follows. Overall, humans and rats show a mild correspondence in terms of which image pairs are more difficult, with a human–rat correlation of 0.18 across all image pairs of the nine test protocols (p<0.001 with permutation test). Albeit significant, this correlation is clearly lower than the maximum value that could be obtained given the reliability of the data. The split-half reliability of the human data was 0.46, corresponding to a full-set reliability of 0.63. We reported above that full-set reliability is 0.58 for the rat data, resulting in a combined reliability of 0.60 (calculated as described in Op de Beeck et al., 2008). Thus, after taking data reliability into account there remains a pronounced discrepancy between rats and humans in terms of how performance varies across image pairs.

The main question of this study is how this discrepancy relates to computationally informed strategies. If we take a closer look specifically at the two CNN-informed test protocols (zero vs. high and high vs. zero), we see an opposite behaviour between animals and humans. Humans performed significantly better in the zero vs. high protocol, that is, where we used stimuli where the earlier layers of the network perform worse than the higher layers, than in the high vs. zero protocol (paired t-test: t(44) = 2.85, p=0.0067). Rats, however, show the opposite (see above for statistics). There even is a significant interaction between species and test protocol (unpaired t-test: t(54) = 2.50, p=0.016). This suggests a different strategy between animals and humans: rats use strategies that are captured in the lower layers of the network, and thus correspond more to low-level visual processing. Humans, however, tend to rely more on strategies captured by the higher layers of the network, and thus we are looking at more high-level visual processing.

As a next step, we calculated the correlation between the generalization across image pairs between the CNN classifier and the human performance of all nine test protocols in an identical manner as for the rat performance (Figure 4). The results are displayed in Figure 6. Overall, we see quite high correlations, especially in the higher layers. This pattern across layers is very different from the pattern in rats where the highest layers showed no correlations, which again suggests that, despite successful generalization, rats rely on decisively lower-level strategies than humans in the same discrimination task.

Figure 6

Download asset Open asset

A multiple linear regression was calculated in an identical manner as we did with the rat performance. The results of this regression indicate a significant effect of the classification scores (F(287,273) = 6.8, p<0.0001, R² = 0.25). Further investigating the 13 predictors showed that in particular the fully connected layers 11–13 of the network were strong predictors in the regression model (see for results Supplementary file 1e of the regression model).

Given the high correlations with fully connected layers, we would expect to not find much evidence for an influence of basic visual dimensions such as brightness and pixel-based similarity. Indeed, we find small correlations with variation in human behavioural performance across image pairs for pixel similarity (0.12) and brightness (–0.12).

In this study, we trained and tested rats and humans in a discrimination task using two-dimensional stimuli, with the two dimensions being concavity and alignment. We tested generalization across a range of viewing conditions. For the last two testing protocols, we used a computational approach to select the stimuli in terms of specifically dissociating low and high stages of processing. Rats were able to learn both dimensions (concavity and alignment) and showed a preference for concavity. Their performance on the testing protocols revealed a wide variety in percentage correct: for some test protocols, they performed just above chance level, for example, zero vs. high, whereas for others they could easily reach about 70% correct (position). Humans, on the other hand, performed better overall, with performances of 80% or higher on the testing protocols. Addressing the question of the complexity of the underlying strategies, rats performed best on the test protocol designed to specifically target lower levels of processing, whereas humans performed best on the high-level processing protocol. Likewise, direct comparisons with artificial neural network layers showed that the variation of rat performance across images was best explained by late convolutional layers, whereas human performance was most associated with representations in fully connected layers.

All animals started by being trained in three training protocols. The first training protocol only included one image pair, the base pair, containing the most different target and distractor without any further transformations. Learning of the individual dimensions of concavity and alignment was investigated through the dimension learning protocol. The results from this dimension learning protocol indicate that our rats have more difficulties learning the alignment dimension as opposed to the concavity dimension. One possible explanation for the superior performance on the concavity dimension could be that the animals were partially solving the task such that the brighter stimulus, that is, the convex base shape, is the distractor and that their strategy is to pick the stimulus with the lowest brightness. This was confirmed by analyses on the third training protocol (transformations) that included small transformations along various dimensions. Nevertheless, the rats still performed above-chance level for trials in which the brightness differences were reversed, indicating that other dimensions are involved and overrule a contribution from brightness. Similar findings have been obtained in human behaviour and neuroscience. For example, despite the clear category selectivity in regions such as the fusiform face area, the selectivity in these regions is also modulated very strongly by various low-level dimensions (Yue et al., 2011). With regard to the size and position transformations, it is important to keep in mind that the animals were freely moving in the touchscreen chambers, and so even for the original base pair were already undergoing changes in retinal size and retinal position. What we manipulate is rather the size and position relative to the rest of the set-up (e.g. relative to screen position and size).

After these three training protocols, the animals were tested for generalization in a variety of testing protocols, each testing a separate transformation on the stimuli. The first six test protocols included rotation along all the three axes, size, position, and light location, followed by a test protocol in which we combined the rotation along the three axes. Overall, we found that the performance of the animals on these test protocols is affected by these transformations, but still significantly above chance in each protocol. Studies in the literature would often stop here or proceed by systematically testing even larger transformations. Stimulus choices are based upon intuitions of what strategy animals might be using and upon theories of how visual perception works. However, in some cases, a further computational modelling of the task and stimuli finds that what intuitively seems like a task of a particular complexity might not be so complex after all. The first tests of invariant object recognition seemed impressive, but were found to be easily solved with earlier layers of processing (Minini and Jeffery, 2006; Vinken and Op de Beeck, 2021). This was recently also highlighted by relatively simple pixel-based analyses (Kell et al., 2020). As another example, Vinken and Op de Beeck, 2021 have used a computational approach to further investigate the levels of information processing in rodents by comparing three hallmark studies that provided evidence for higher order visual processing in rodents (Djurdjevic et al., 2018; Vinken et al., 2014; Zoccolan et al., 2009) with CNNs. They found that for all three studies the low- and mid-level layers captured the rat performances best, providing thus evidence against the previously concluded high-level visual processing in rodents.

For these reasons, we decided to directly test image pairs through computational modelling with CNNs and select pairs that are particularly suited for dissociating different levels of processing. Stimuli were chosen by a CNN from a very large set of possible stimuli and combinations, such that the higher layers and the lower layers of the network make distinct errors on classifying the stimuli (zero vs. high and high vs. zero protocol), and thus are diagnostic of the level of underlying visual strategies. We chose to work with Alexnet as this is a network that has been used as a benchmark in many previous studies (e.g. Cadieu et al., 2014; Groen et al., 2018; Kalfas et al., 2018; Nayebi et al., 2023; Zeman et al., 2020), including studies that used more complex stimuli than the stimulus space in this study. The stimuli of the zero vs. high protocol included stimuli where the higher layers of the network performed better than the lower layers, and thus they address higher-level visual processing. The opposite can be said for the high vs. zero protocol, which includes stimuli that specifically target lower-level visual processing, given that the lower layers of the network perform best on these stimuli. After presenting these stimuli to the animals, we found that our rats performed best in the high vs. zero protocol, suggesting that they focus on low-level visual cues to solve this discrimination task. We found the opposite CNN pattern for humans, indicating that they use high-level visual processing. These findings provide more direct information about the level of processing that underlies the behavioural strategies compared to overall performance or to effects of image manipulations.

This is a new promising way to design experiments in a way that is computationally informed rather than based on researcher intuitions or qualitative predictions. It is in line with the literature that a typical deep neural network, AlexNet and also more complex ones, can explain human and animal behaviour to a certain extent but not fully. The explained variance might differ among CNNs, and there might be CNNs that can explain a higher proportion of rat or human behaviour. Most relevant for this study is that CNNs tend to agree in terms of how representations change from lower to higher hierarchical layers because this is the transformation that we have targeted in the zero vs. high and high vs. zero testing protocols. Pinto et al., 2008 already revealed that a simple V1-like model can sometimes result in surprisingly good object recognition performance. This aspect of our findings is also in line with the observation of Vinken and Op de Beeck, 2021 that the performance of rats in many previous tasks might not be indicative of highly complex representations. Nevertheless, there is still a relative difference in complexity between lower and higher levels in the hierarchy. That is what we capitalize upon with the zero vs. high and high vs. zero protocols. Thus, it might be more fruitful to explicitly contrast different levels of processing in a relative way rather than trying to pinpoint behaviour to specific levels of processing.

Partially thanks to these computationally inspired tests, our total dataset finds a marked dissociation between how humans and rats solve this object recognition task. Even in the sessions where only the old pairs are shown, the animals performed lower than humans. This was most likely due to motivation and/or distractibility. Our analyses show dissociation between humans and rats most convincingly by correlating the variation in performance across image trials with the predictions of CNN layers. There were significant correlations with multiple layers in both species. In humans, the most pronounced correlations were present for the highest, fully connected layers, while in rats correlations were limited to low and middle convolutional layers. This is the most direct evidence available in the literature that rats resolve object recognition tasks through a very different and computationally simpler strategy compared to humans. The CNN approach does not inform us how we can verbalize this simpler strategy, but based upon earlier work (Schnell et al., 2023; Vermaercke and Op de Beeck, 2012) we would hypothesize that rats rely upon visual contrast features (e.g. this area is darker/lighter than that other area). Such contrast features are also used by humans and monkeys, for example, for face detection (Ohayon et al., 2012; Sinha, 2002), but in addition humans have access to more complex strategies that, for example, refer to complex shape features such as aspect ratio and symmetry (Bossens and Op de Beeck, 2016). Tests in this study reveal that other features that partially explain rat performance include basic dimensions such as brightness and pixel-based similarity, the latter being a proxy for retinotopically-based computations that are expected to be present in convolutional layers.

Our analyses of rat behaviour and DNN modelling do not take into account potential trial-to-trial variability in the distance and position of the rat’s head. From earlier work we can derive that rats typically make their decision from about 12 cm from the stimulus display (Crijns and Op de Beeck, 2019), but we have no information on trial-to-trial variability. We can hypothesize about the possible effect. If such variability would exist, then it would artificially increase the variation in distance and position during training, and thus help the animals achieve higher levels of invariance during testing. As a consequence, the difference between rat and human performance in terms of inferred level of processing might even increase under more controlled circumstances.

For future studies, it will be highly valuable to use this computational informed strategy on a wider battery of behavioural tasks, as well as a wider range of species such as tree shrews and marmosets (Callahan and Petry, 2000; Kell et al., 2020; Kell et al., 2021; Meyer et al., 2022; Petry et al., 2012; Petry and Bickford, 2019). One step further, we can use the information from computational modelling together with behaviour and how it differs among stimuli to further select stimuli for neurophysiological investigations of neuronal response properties along the visual information processing hierarchy, in this way following experimental designs that are optimized for highlighting the primary differences between processing stages and between species.

The data has been made publicly available via the Open Science Framework and can be accessed at https://osf.io/9eqyz/.

1. 1. Schnell AE
   2. Op de Beeck H
   3. Vinken K
   4. Leemans M

   (2023) Open Science Framework

   Multidimensional stimulus set.

   https://doi.org/10.17605/OSF.IO/V9WBN
2. 1. Schnell AE
   2. Op de Beeck H
   3. Vinken K
   4. Leemans M

   (2023) Open Science Framework

   A computationally informed comparison between the strategies of rodents and humans in visual object recognition.

   https://doi.org/10.17605/OSF.IO/9EQYZ

1. 1. Alemi-Neissi A
   2. Rosselli FB
   3. Zoccolan D

   (2013) Multifeatural shape processing in rats engaged in invariant visual object recognition

   The Journal of Neuroscience 33:5939–5956.

   https://doi.org/10.1523/JNEUROSCI.3629-12.2013

   • PubMed
   • Google Scholar
2. 1. Avberšek LK
   2. Zeman A
   3. Op de Beeck H

   (2021) Training for object recognition with increasing spatial frequency: A comparison of deep learning with human vision

   Journal of Vision 21:14.

   https://doi.org/10.1167/jov.21.10.14

   • PubMed
   • Google Scholar
3. 1. Bossens C
   2. Op de Beeck HP

   (2016) Linear and non-linear visual feature learning in rat and humans

   Frontiers in Behavioral Neuroscience 10:235.

   https://doi.org/10.3389/fnbeh.2016.00235

   • PubMed
   • Google Scholar
4. 1. Cadieu CF
   2. Hong H
   3. Yamins DLK
   4. Pinto N
   5. Ardila D
   6. Solomon EA
   7. Majaj NJ
   8. DiCarlo JJ

   (2014) Deep neural networks rival the representation of primate IT cortex for core visual object recognition

   PLOS Computational Biology 10:e1003963.

   https://doi.org/10.1371/journal.pcbi.1003963

   • PubMed
   • Google Scholar
5. 1. Callahan TL
   2. Petry HM

   (2000) Psychophysical measurement of temporal modulation sensitivity in the tree shrew (Tupaia belangeri)

   Vision Research 40:455–458.

   https://doi.org/10.1016/s0042-6989(99)00194-7

   • PubMed
   • Google Scholar
6. 1. Crijns E
   2. Op de Beeck H

   (2019) The visual acuity of rats in touchscreen setups

   Vision 4:4.

   https://doi.org/10.3390/vision4010004

   • PubMed
   • Google Scholar
7. 1. De Keyser R
   2. Bossens C
   3. Kubilius J
   4. Op de Beeck HP

   (2015) Cue-invariant shape recognition in rats as tested with second-order contours

   Journal of Vision 15:14.

   https://doi.org/10.1167/15.15.14

   • PubMed
   • Google Scholar
8. 1. Djurdjevic V
   2. Ansuini A
   3. Bertolini D
   4. Macke JH
   5. Zoccolan D

   (2018) Accuracy of rats in discriminating visual objects is explained by the complexity of their perceptual strategy

   Current Biology 28:1005–1015.

   https://doi.org/10.1016/j.cub.2018.02.037

   • PubMed
   • Google Scholar
9. 1. Duyck S
   2. Martens F
   3. Chen CY
   4. Op de Beeck H

   (2021) How visual expertise changes representational geometry: A behavioral and neural perspective

   Journal of Cognitive Neuroscience 33:2461–2476.

   https://doi.org/10.1162/jocn_a_01778

   • PubMed
   • Google Scholar
10. 1. Groen II
    2. Greene MR
    3. Baldassano C
    4. Fei-Fei L
    5. Beck DM
    6. Baker CI

    (2018) Distinct contributions of functional and deep neural network features to representational similarity of scenes in human brain and behavior

    eLife 7:e32962.

    https://doi.org/10.7554/eLife.32962

    • PubMed
    • Google Scholar
11. 1. Güçlü U
    2. van Gerven MAJ

    (2015) Deep neural networks reveal a gradient in the complexity of neural representations across the ventral stream

    The Journal of Neuroscience 35:10005–10014.

    https://doi.org/10.1523/JNEUROSCI.5023-14.2015

    • PubMed
    • Google Scholar
12. 1. Kalfas I
    2. Vinken K
    3. Vogels R

    (2018) Representations of regular and irregular shapes by deep Convolutional Neural Networks, monkey inferotemporal neurons and human judgments

    PLOS Computational Biology 14:e1006557.

    https://doi.org/10.1371/journal.pcbi.1006557

    • PubMed
    • Google Scholar
13. 1. Kar K
    2. Kubilius J
    3. Schmidt K
    4. Issa EB
    5. DiCarlo JJ

    (2019) Evidence that recurrent circuits are critical to the ventral stream’s execution of core object recognition behavior

    Nature Neuroscience 22:974–983.

    https://doi.org/10.1038/s41593-019-0392-5

    • PubMed
    • Google Scholar
14. Preprint

    1. Kell AJE
    2. Bokor SL
    3. Jeon YN
    4. Toosi T
    5. Issa EB

    (2020) Conserved Core Visual Object Recognition across Simian Primates: Marmoset Image-by-Image Behavior Mirrors That of Humans and Macaques

    bioRxiv.

    https://doi.org/10.1101/2020.10.19.345561

    • Google Scholar
15. Preprint

    1. Kell AJE
    2. Bokor SL
    3. Jeon YN
    4. Toosi T
    5. Issa EB

    (2021) Brain Organization, Not Size Alone, as Key to High-Level Vision: Evidence from Marmoset Monkeys

    bioRxiv.

    https://doi.org/10.1101/2020.10.19.345561v3

    • Google Scholar
16. 1. Kubilius J
    2. Bracci S
    3. Op de Beeck HP

    (2016) Deep neural networks as a computational model for human shape sensitivity

    PLOS Computational Biology 12:e1004896.

    https://doi.org/10.1371/journal.pcbi.1004896

    • PubMed
    • Google Scholar
17. 1. Logothetis NK
    2. Sheinberg DL

    (1996) Visual object recognition

    Annual Review of Neuroscience 19:577–621.

    https://doi.org/10.1146/annurev.ne.19.030196.003045

    • PubMed
    • Google Scholar
18. 1. Matteucci G
    2. Bellacosa Marotti R
    3. Riggi M
    4. Rosselli FB
    5. Zoccolan D

    (2019) Nonlinear processing of shape information in rat lateral extrastriate cortex

    The Journal of Neuroscience 39:1649–1670.

    https://doi.org/10.1523/JNEUROSCI.1938-18.2018

    • PubMed
    • Google Scholar
19. Conference

    1. Meyer EE
    2. Ong WS
    3. Balboa M
    4. Arcaro MJ

    (2022)

    Assessing tree shrew high-level visual behavior using conventional and natural paradigms [Poster presentation]

    Society for Neuroscience, November 12-16.

    • Google Scholar
20. 1. Minini L
    2. Jeffery KJ

    (2006) Do rats use shape to solve “shape discriminations”?

    Learning & Memory 13:287–297.

    https://doi.org/10.1101/lm.84406

    • PubMed
    • Google Scholar
21. Preprint

    1. Nayebi A
    2. Kong NCL
    3. Zhuang C
    4. Gardner JL
    5. Norcia AM
    6. Yamins DLK

    (2023) Mouse Visual Cortex as a Limited Resource System That Self-Learns an Ecologically-General Representation

    bioRxiv.

    https://doi.org/10.1101/2021.06.16.448730

    • Google Scholar
22. 1. Ohayon S
    2. Freiwald WA
    3. Tsao DY

    (2012) What makes a cell face selective

    The Importance of Contrast. Neuron 74:567–581.

    https://doi.org/10.1016/j.neuron.2012.03.024

    • Google Scholar
23. 1. Op de Beeck HP
    2. Torfs K
    3. Wagemans J

    (2008) Perceived shape similarity among unfamiliar objects and the organization of the human object vision pathway

    The Journal of Neuroscience 28:10111–10123.

    https://doi.org/10.1523/JNEUROSCI.2511-08.2008

    • PubMed
    • Google Scholar
24. 1. Petry HM
    2. Clark C
    3. Day-Brown J
    4. Bolin RT
    5. Bickford M

    (2012) Behavioral measurement of RDK velocity discrimination thresholds in the tree shrew

    Journal of Vision 12:1223.

    https://doi.org/10.1167/12.9.1223

    • Google Scholar
25. 1. Petry HM
    2. Bickford ME

    (2019) The second visual system of the tree shrew

    The Journal of Comparative Neurology 527:679–693.

    https://doi.org/10.1002/cne.24413

    • PubMed
    • Google Scholar
26. 1. Pinto N
    2. Cox DD
    3. DiCarlo JJ

    (2008) Why is real-world visual object recognition hard?

    PLOS Computational Biology 4:e27.

    https://doi.org/10.1371/journal.pcbi.0040027

    • PubMed
    • Google Scholar
27. 1. Pospisil DA
    2. Pasupathy A
    3. Bair W

    (2018) “Artiphysiology” reveals V4-like shape tuning in a deep network trained for image classification

    eLife 7:e38242.

    https://doi.org/10.7554/eLife.38242

    • PubMed
    • Google Scholar
28. 1. Schnell AE
    2. Van den Bergh G
    3. Vermaercke B
    4. Gijbels K
    5. Bossens C
    6. de Beeck HO

    (2019) Face categorization and behavioral templates in rats

    Journal of Vision 19:9.

    https://doi.org/10.1167/19.14.9

    • PubMed
    • Google Scholar
29. 1. Schnell AE
    2. Vinken K
    3. de Beeck HO

    (2023) The importance of contrast features in rat vision

    Scientific Reports 13:459.

    https://doi.org/10.1038/s41598-023-27533-3

    • PubMed
    • Google Scholar
30. Book

    1. Sinha P

    (2002) Qualitative representations for recognition

    In: Bülthoff HH, Wallraven C, Lee SW, Poggio TA, editors. Biologically Motivated Computer Vision. Springer. pp. 249–262.

    https://doi.org/10.1007/3-540-36181-2

    • Google Scholar
31. 1. Tafazoli S
    2. Di Filippo A
    3. Zoccolan D

    (2012) Transformation-tolerant object recognition in rats revealed by visual priming

    The Journal of Neuroscience 32:21–34.

    https://doi.org/10.1523/JNEUROSCI.3932-11.2012

    • PubMed
    • Google Scholar
32. 1. Tafazoli S
    2. Safaai H
    3. De Franceschi G
    4. Rosselli FB
    5. Vanzella W
    6. Riggi M
    7. Buffolo F
    8. Panzeri S
    9. Zoccolan D

    (2017) Emergence of transformation-tolerant representations of visual objects in rat lateral extrastriate cortex

    eLife 6:e22794.

    https://doi.org/10.7554/eLife.22794

    • PubMed
    • Google Scholar
33. 1. Vermaercke B
    2. Op de Beeck HP

    (2012) A multivariate approach reveals the behavioral templates underlying visual discrimination in rats

    Current Biology 22:50–55.

    https://doi.org/10.1016/j.cub.2011.11.041

    • PubMed
    • Google Scholar
34. 1. Vermaercke B
    2. Gerich FJ
    3. Ytebrouck E
    4. Arckens L
    5. Op de Beeck HP
    6. Van den Bergh G

    (2014) Functional specialization in rat occipital and temporal visual cortex

    Journal of Neurophysiology 112:1963–1983.

    https://doi.org/10.1152/jn.00737.2013

    • PubMed
    • Google Scholar
35. 1. Vinken K
    2. Vermaercke B
    3. Op de Beeck HP

    (2014) Visual categorization of natural movies by rats

    The Journal of Neuroscience 34:10645–10658.

    https://doi.org/10.1523/JNEUROSCI.3663-13.2014

    • PubMed
    • Google Scholar
36. 1. Vinken K
    2. Van den Bergh G
    3. Vermaercke B
    4. Op de Beeck HP

    (2016) Neural representations of natural and scrambled movies progressively change from rat striate to temporal cortex

    Cerebral Cortex 26:3310–3322.

    https://doi.org/10.1093/cercor/bhw111

    • PubMed
    • Google Scholar
37. 1. Vinken K
    2. Op de Beeck H

    (2021) Using deep neural networks to evaluate object vision tasks in rats

    PLOS Computational Biology 17:e1008714.

    https://doi.org/10.1371/journal.pcbi.1008714

    • PubMed
    • Google Scholar
38. 1. Yue X
    2. Cassidy BS
    3. Devaney KJ
    4. Holt DJ
    5. Tootell RBH

    (2011) Lower-level stimulus features strongly influence responses in the fusiform face area

    Cerebral Cortex 21:35–47.

    https://doi.org/10.1093/cercor/bhq050

    • PubMed
    • Google Scholar
39. 1. Zeman AA
    2. Ritchie JB
    3. Bracci S
    4. Op de Beeck H

    (2020) Orthogonal representations of object shape and category in deep convolutional neural networks and human visual cortex

    Scientific Reports 10:2453.

    https://doi.org/10.1038/s41598-020-59175-0

    • PubMed
    • Google Scholar
40. 1. Zoccolan D
    2. Oertelt N
    3. DiCarlo JJ
    4. Cox DD

    (2009) A rodent model for the study of invariant visual object recognition

    PNAS 106:8748–8753.

    https://doi.org/10.1073/pnas.0811583106

    • PubMed
    • Google Scholar
41. 1. Zoccolan D

    (2015) Invariant visual object recognition and shape processing in rats

    Behavioural Brain Research 285:10–33.

    https://doi.org/10.1016/j.bbr.2014.12.053

    • PubMed
    • Google Scholar

Author details

1. Anna Elisabeth Schnell

   Department of Brain and Cognition & Leuven Brain Institute, Leuven, Belgium

   Contribution

   Conceptualization, Data curation, Formal analysis, Investigation, Visualization, Methodology, Writing - original draft, Writing - review and editing

   For correspondence

   annaelisabeth.schnell@kuleuven.be

   Competing interests

   No competing interests declared

   "This ORCID iD identifies the author of this article:" 0000-0001-9278-1817

2. Maarten Leemans

   Department of Brain and Cognition & Leuven Brain Institute, Leuven, Belgium

   Contribution

   Writing - review and editing

   Competing interests

   No competing interests declared

3. Kasper Vinken

   Department of Neurobiology, Harvard Medical School, Boston, United States

   Contribution

   Writing - review and editing

   Competing interests

   No competing interests declared

   "This ORCID iD identifies the author of this article:" 0000-0001-7038-9638

4. Hans Op de Beeck

   Department of Brain and Cognition & Leuven Brain Institute, Leuven, Belgium

   Contribution

   Conceptualization, Data curation, Supervision, Funding acquisition, Investigation, Methodology, Project administration, Writing - review and editing

   Competing interests

   No competing interests declared

• 587

  views

• 43

  downloads

• 0

  citations

Views, downloads and citations are aggregated across all versions of this paper published by eLife.