When the ventral visual stream is not enough: A deep learning account of medial temporal lobe involvement in perception

Experience transforms crossmodal object representations in the anterior temporal lobes

Combining information from multiple senses is essential to object recognition, core to the ability to learn concepts, make new inferences, and generalize across distinct entities. Yet how the mind combines sensory input into coherent crossmodal representations - the crossmodal binding problem - remains poorly understood. Here, we applied multi-echo fMRI across a 4-day paradigm, in which participants learned three-dimensional crossmodal representations created from well-characterized unimodal visual shape and sound features. Our novel paradigm decoupled the learned crossmodal object representations from their baseline unimodal shapes and sounds, thus allowing us to track the emergence of crossmodal object representations as they were learned by healthy adults. Critically, we found that two anterior temporal lobe structures - temporal pole and perirhinal cortex - differentiated learned from non-learned crossmodal objects, even when controlling for the unimodal features that composed those objects. These results provide evidence for integrated crossmodal object representations in the anterior temporal lobes that were different from the representations for the unimodal features. Furthermore, we found that perirhinal cortex representations were by default biased toward visual shape, but this initial visual bias was attenuated by crossmodal learning. Thus, crossmodal learning transformed perirhinal representations such that they were no longer predominantly grounded in the visual modality, which may be a mechanism by which object concepts gain their abstraction.

Encoding of Visual Objects in the Human Medial Temporal Lobe

The human medial temporal lobe (MTL) plays a crucial role in recognizing visual objects, a key cognitive function that relies on the formation of semantic representations. Nonetheless, it remains unknown how visual information of general objects is translated into semantic representations in the MTL. Furthermore, the debate about whether the human MTL is involved in perception has endured for a long time. To address these questions, we investigated three distinct models of neural object coding-semantic coding, axis-based feature coding, and region-based feature coding-in each subregion of the human MTL, using high-resolution fMRI in two male and six female participants. Our findings revealed the presence of semantic coding throughout the MTL, with a higher prevalence observed in the parahippocampal cortex (PHC) and perirhinal cortex (PRC), while axis coding and region coding were primarily observed in the earlier regions of the MTL. Moreover, we demonstrated that voxels exhibiting axis coding supported the transition to region coding and contained information relevant to semantic coding. Together, by providing a detailed characterization of neural object coding schemes and offering a comprehensive summary of visual coding information for each MTL subregion, our results not only emphasize a clear role of the MTL in perceptual processing but also shed light on the translation of perception-driven representations of visual features into memory-driven representations of semantics along the MTL processing pathway.

Emergence of Emotion Selectivity in Deep Neural Networks Trained to Recognize Visual Objects

Recent neuroimaging studies have shown that the visual cortex plays an important role in representing the affective significance of visual input. The origin of these affect-specific visual representations is debated: they are intrinsic to the visual system versus they arise through reentry from frontal emotion processing structures such as the amygdala. We examined this problem by combining convolutional neural network (CNN) models of the human ventral visual cortex pre-trained on ImageNet with two datasets of affective images. Our results show that in all layers of the CNN models, there were artificial neurons that responded consistently and selectively to neutral, pleasant, or unpleasant images and lesioning these neurons by setting their output to zero or enhancing these neurons by increasing their gain led to decreased or increased emotion recognition performance respectively. These results support the idea that the visual system may have the intrinsic ability to represent the affective significance of visual input and suggest that CNNs offer a fruitful platform for testing neuroscientific theories.

Using deep neural networks to disentangle visual and semantic information in human perception and memory

Mental representations of familiar categories are composed of visual and semantic information. Disentangling the contributions of visual and semantic information in humans is challenging because they are intermixed in mental representations. Deep neural networks that are trained either on images or on text or by pairing images and text enable us now to disentangle human mental representations into their visual, visual-semantic and semantic components. Here we used these deep neural networks to uncover the content of human mental representations of familiar faces and objects when they are viewed or recalled from memory. The results show a larger visual than semantic contribution when images are viewed and a reversed pattern when they are recalled. We further reveal a previously unknown unique contribution of an integrated visual-semantic representation in both perception and memory. We propose a new framework in which visual and semantic information contribute independently and interactively to mental representations in perception and memory.

Grounding neuroscience in behavioral changes using artificial neural networks

Connecting neural activity to function is a common aim in neuroscience. How to define and conceptualize function, however, can vary. Here I focus on grounding this goal in the specific question of how a given change in behavior is produced by a change in neural circuits or activity. Artificial neural network models offer a particularly fruitful format for tackling such questions because they use neural mechanisms to perform complex transformations and produce appropriate behavior. Therefore, they can be a means of causally testing the extent to which a neural change can be responsible for an experimentally observed behavioral change. Furthermore, because the field of interpretability in artificial intelligence has similar aims, neuroscientists can look to interpretability methods for new ways of identifying neural features that drive performance and behaviors.

Computational models can distinguish the contribution from different mechanisms to familiarity recognition

Familiarity is the strange feeling of knowing that something has already been seen in our past. Over the past decades, several attempts have been made to model familiarity using artificial neural networks. Recently, two learning algorithms successfully reproduced the functioning of the perirhinal cortex, a key structure involved during familiarity: Hebbian and anti-Hebbian learning. However, performance of these learning rules is very different from one to another thus raising the question of their complementarity. In this work, we designed two distinct computational models that combined Deep Learning and a Hebbian learning rule to reproduce familiarity on natural images, the Hebbian model and the anti-Hebbian model, respectively. We compared the performance of both models during different simulations to highlight the inner functioning of both learning rules. We showed that the anti-Hebbian model fits human behavioral data whereas the Hebbian model fails to fit the data under large training set sizes. Besides, we observed that only our Hebbian model is highly sensitive to homogeneity between images. Taken together, we interpreted these results considering the distinction between absolute and relative familiarity. With our framework, we proposed a novel way to distinguish the contribution of these familiarity mechanisms to the overall feeling of familiarity. By viewing them as complementary, our two models allow us to make new testable predictions that could be of interest to shed light on the familiarity phenomenon.

Recurrent connectivity supports higher-level visual and semantic object representations in the brain

Visual object recognition has been traditionally conceptualised as a predominantly feedforward process through the ventral visual pathway. While feedforward artificial neural networks (ANNs) can achieve human-level classification on some image-labelling tasks, it's unclear whether computational models of vision alone can accurately capture the evolving spatiotemporal neural dynamics. Here, we probe these dynamics using a combination of representational similarity and connectivity analyses of fMRI and MEG data recorded during the recognition of familiar, unambiguous objects. Modelling the visual and semantic properties of our stimuli using an artificial neural network as well as a semantic feature model, we find that unique aspects of the neural architecture and connectivity dynamics relate to visual and semantic object properties. Critically, we show that recurrent processing between the anterior and posterior ventral temporal cortex relates to higher-level visual properties prior to semantic object properties, in addition to semantic-related feedback from the frontal lobe to the ventral temporal lobe between 250 and 500 ms after stimulus onset. These results demonstrate the distinct contributions made by semantic object properties in explaining neural activity and connectivity, highlighting it as a core part of object recognition not fully accounted for by current biologically inspired neural networks.

Functional neuroimaging as a catalyst for integrated neuroscience

Functional magnetic resonance imaging (fMRI) enables non-invasive access to the awake, behaving human brain. By tracking whole-brain signals across a diverse range of cognitive and behavioural states or mapping differences associated with specific traits or clinical conditions, fMRI has advanced our understanding of brain function and its links to both normal and atypical behaviour. Despite this headway, progress in human cognitive neuroscience that uses fMRI has been relatively isolated from rapid advances in other subdomains of neuroscience, which themselves are also somewhat siloed from one another. In this Perspective, we argue that fMRI is well-placed to integrate the diverse subfields of systems, cognitive, computational and clinical neuroscience. We first summarize the strengths and weaknesses of fMRI as an imaging tool, then highlight examples of studies that have successfully used fMRI in each subdomain of neuroscience. We then provide a roadmap for the future advances that will be needed to realize this integrative vision. In this way, we hope to demonstrate how fMRI can help usher in a new era of interdisciplinary coherence in neuroscience.

Exploring strategy differences between humans and monkeys with recurrent neural networks

Animal models are used to understand principles of human biology. Within cognitive neuroscience, non-human primates are considered the premier model for studying decision-making behaviors in which direct manipulation experiments are still possible. Some prominent studies have brought to light major discrepancies between monkey and human cognition, highlighting problems with unverified extrapolation from monkey to human. Here, we use a parallel model system-artificial neural networks (ANNs)-to investigate a well-established discrepancy identified between monkeys and humans with a working memory task, in which monkeys appear to use a recency-based strategy while humans use a target-selective strategy. We find that ANNs trained on the same task exhibit a progression of behavior from random behavior (untrained) to recency-like behavior (partially trained) and finally to selective behavior (further trained), suggesting monkeys and humans may occupy different points in the same overall learning progression. Surprisingly, what appears to be recency-like behavior in the ANN, is in fact an emergent non-recency-based property of the organization of the neural network's state space during its development through training. We find that explicit encouragement of recency behavior during training has a dual effect, not only causing an accentuated recency-like behavior, but also speeding up the learning process altogether, resulting in an efficient shaping mechanism to achieve the optimal strategy. Our results suggest a new explanation for the discrepency observed between monkeys and humans and reveal that what can appear to be a recency-based strategy in some cases may not be recency at all.

When the whole is only the parts: non-holistic object parts predominate face-cell responses to illusory faces

Humans are inclined to perceive faces in everyday objects with a face-like configuration. This illusion, known as face pareidolia, is often attributed to a specialized network of 'face cells' in primates. We found that face cells in macaque inferotemporal cortex responded selectively to pareidolia images, but this selectivity did not require a holistic, face-like configuration, nor did it encode human faceness ratings. Instead, it was driven mostly by isolated object parts that are perceived as eyes only within a face-like context. These object parts lack usual characteristics of primate eyes, pointing to the role of lower-level features. Our results suggest that face-cell responses are dominated by local, generic features, unlike primate visual perception, which requires holistic information. These findings caution against interpreting neural activity through the lens of human perception. Doing so could impose human perceptual biases, like seeing faces where none exist, onto our understanding of neural activity.

Perception and Memory in the Ventral Visual Stream and Medial Temporal Lobe

Perception and memory are traditionally thought of as separate cognitive functions, supported by distinct brain regions. The canonical perspective is that perceptual processing of visual information is supported by the ventral visual stream, whereas long-term declarative memory is supported by the medial temporal lobe. However, this modular framework cannot account for the increasingly large body of evidence that reveals a role for early visual areas in long-term recognition memory and a role for medial temporal lobe structures in high-level perceptual processing. In this article, we review relevant research conducted in humans, nonhuman primates, and rodents. We conclude that the evidence is largely inconsistent with theoretical proposals that draw sharp functional boundaries between perceptual and memory systems in the brain. Instead, the weight of the empirical findings is best captured by a representational-hierarchical model that emphasizes differences in content, rather than in cognitive processes within the ventral visual stream and medial temporal lobe.

Real-time simulations of ADF STEM probe position-integrated scattering cross-sections for single element fcc crystals in zone axis orientation using a densely connected neural network

Quantification of annular dark field (ADF) scanning transmission electron microscopy (STEM) images in terms of composition or thickness often relies on probe-position integrated scattering cross sections (PPISCS). In order to compare experimental PPISCS with theoretically predicted ones, expensive simulations are needed for a given specimen, zone axis orientation, and a variety of microscope settings. The computation time of such simulations can be in the order of hours using a single GPU card. ADF STEM simulations can be efficiently parallelized using multiple GPUs, as the calculation of each pixel is independent of other pixels. However, most research groups do not have the necessary hardware, and, in the best-case scenario, the simulation time will only be reduced proportionally to the number of GPUs used. In this manuscript, we use a learning approach and present a densely connected neural network that is able to perform real-time ADF STEM PPISCS predictions as a function of atomic column thickness for most common face-centered cubic (fcc) crystals (i.e., Al, Cu, Pd, Ag, Pt, Au and Pb) along [100] and [111] zone axis orientations, root-mean-square displacements, and microscope parameters. The proposed architecture is parameter efficient and yields accurate predictions for the PPISCS values for a wide range of input parameters that are commonly used for aberration-corrected transmission electron microscopes.

Patterns of Impaired Neurocognitive Performance on the Global Neuropsychological Assessment, and Their Brain Structural Correlates in Recent-onset and Chronic Schizophrenia

Objective

Schizophrenia is associated with impairment in multiple cognitive domains. There is a paucity of research on the effect of prolonged illness duration (≥ 15 years) on cognitive performance along multiple domains. In this pilot study, we used the Global Neuropsychological Assessment (GNA), a brief cognitive battery, to explore the patterns of cognitive impairment in recent-onset (≤ 2 years) compared to chronic schizophrenia (≥ 15 years), and correlate cognitive performance with brain morphometry in patients and healthy adults.

Methods

We assessed cognitive performance in patients with recent-onset (n = 17, illness duration ≤ 2 years) and chronic schizophrenia (n = 14, duration ≥ 15 years), and healthy adults (n = 16) using the GNA and examined correlations between cognitive scores and gray matter volumes computed from T1-weighted magnetic resonance imaging images.

Results

We observed cognitive deficits affecting multiple domains in the schizophrenia samples. Selectively greater impairment of perceptual comparison speed was found in adults with chronic schizophrenia (p = 0.009, η2partial = 0.25). In the full sample (n = 47), perceptual comparison speed correlated significantly with gray matter volumes in the anterior and medial temporal lobes (TFCE, FWE p < 0.01).

Conclusion

Along with generalized deficit across multiple cognitive domains, selectively greater impairment of perceptual comparison speed appears to characterize chronic schizophrenia. This pattern might indicate an accelerated or premature cognitive aging. Anterior-medial temporal gray matter volumes especially of the left hemisphere might underlie the impairment noted in this domain in schizophrenia.

Neural learning rules for generating flexible predictions and computing the successor representation

The predictive nature of the hippocampus is thought to be useful for memory-guided cognitive behaviors. Inspired by the reinforcement learning literature, this notion has been formalized as a predictive map called the successor representation (SR). The SR captures a number of observations about hippocampal activity. However, the algorithm does not provide a neural mechanism for how such representations arise. Here, we show the dynamics of a recurrent neural network naturally calculate the SR when the synaptic weights match the transition probability matrix. Interestingly, the predictive horizon can be flexibly modulated simply by changing the network gain. We derive simple, biologically plausible learning rules to learn the SR in a recurrent network. We test our model with realistic inputs and match hippocampal data recorded during random foraging. Taken together, our results suggest that the SR is more accessible in neural circuits than previously thought and can support a broad range of cognitive functions.

The entorhinal-DG/CA3 pathway in the medial temporal lobe retains visual working memory of a simple surface feature

Classic models consider working memory (WM) and long-term memory as distinct mental faculties that are supported by different neural mechanisms. Yet, there are significant parallels in the computation that both types of memory require. For instance, the representation of precise item-specific memory requires the separation of overlapping neural representations of similar information. This computation has been referred to as pattern separation, which can be mediated by the entorhinal-DG/CA3 pathway of the medial temporal lobe (MTL) in service of long-term episodic memory. However, although recent evidence has suggested that the MTL is involved in WM, the extent to which the entorhinal-DG/CA3 pathway supports precise item-specific WM has remained elusive. Here, we combine an established orientation WM task with high-resolution fMRI to test the hypothesis that the entorhinal-DG/CA3 pathway retains visual WM of a simple surface feature. Participants were retrospectively cued to retain one of the two studied orientation gratings during a brief delay period and then tried to reproduce the cued orientation as precisely as possible. By modeling the delay-period activity to reconstruct the retained WM content, we found that the anterior-lateral entorhinal cortex (aLEC) and the hippocampal DG/CA3 subfield both contain item-specific WM information that is associated with subsequent recall fidelity. Together, these results highlight the contribution of MTL circuitry to item-specific WM representation.

The medial temporal lobe supports the quality of visual short-term memory representation

The quality of short-term memory (STM) underlies our ability to recall the exact details of a recent event, yet how the human brain enables this core cognitive function remains poorly understood. Here we use multiple experimental approaches to test the hypothesis that the quality of STM, such as its precision or fidelity, relies on the medial temporal lobe (MTL), a region commonly associated with the ability to distinguish similar information remembered in long-term memory. First, with intracranial recordings, we find that delay-period MTL activity retains item-specific STM content that is predictive of subsequent recall precision. Second, STM recall precision is associated with an increase in the strength of intrinsic MTL-to-neocortical functional connections during a brief retention interval. Finally, perturbing the MTL through electrical stimulation or surgical removal can selectively reduce STM precision. Collectively, these findings provide converging evidence that the MTL is critically involved in the quality of STM representation.

Medial temporal lobe structure, mnemonic and perceptual discrimination in healthy older adults and those at risk for mild cognitive impairment

Cognitive tests sensitive to the integrity of the medial temporal lobe (MTL), such as mnemonic discrimination of perceptually similar stimuli, may be useful early markers of risk for cognitive decline in older populations. Perceptual discrimination of stimuli with overlapping features also relies on MTL but remains relatively unexplored in this context. We assessed mnemonic discrimination in two test formats (Forced Choice, Yes/No) and perceptual discrimination of objects and scenes in 111 community-dwelling older adults at different risk status for cognitive impairment based on neuropsychological screening. We also investigated associations between performance and MTL sub-region volume and thickness. The at-risk group exhibited reduced entorhinal thickness and impaired perceptual and mnemonic discrimination. Perceptual discrimination impairment partially explained group differences in mnemonic discrimination and correlated with entorhinal thickness. Executive dysfunction accounted for Yes/No deficits in at-risk adults, demonstrating the importance of test format for the interpretation of memory decline. These results suggest that perceptual discrimination tasks may be useful tools for detecting incipient cognitive impairment related to reduced MTL integrity in nonclinical populations.

Magnocellular and parvocellular contributions to brain network dysfunction during learning and memory: Implications for schizophrenia

Memory deficits are core features of schizophrenia, and a central aim in biological psychiatry is to identify the etiology of these deficits. Scrutiny is naturally focused on the dorsolateral prefrontal cortex and the hippocampal cortices, given these structures' roles in memory and learning. The fronto-hippocampal framework is valuable but restrictive. Network-based underpinnings of learning and memory are substantially diverse and include interactions between hetero-modal and early sensory networks. Thus, a loss of fidelity in sensory information may impact memorial and cognitive processing in higher-order brain sub-networks, becoming a sensory source for learning and memory deficits. In this overview, we suggest that impairments in magno- and parvo-cellular visual pathways result in degraded inputs to core learning and memory networks. The ascending cascade of aberrant neural events significantly contributes to learning and memory deficits in schizophrenia. We outline the network bases of these effects, and suggest that any network perspectives of dysfunction in schizophrenia must assess the impact of impaired perceptual contributions. Finally, we speculate on how this framework enriches the space of biomarkers and expands intervention strategies to ameliorate this prototypical disconnection syndrome.

Prefrontal and Medial Temporal Lobe Cortical Contributions to Visual Short-Term Memory

A number of recent studies have indicated that the medial temporal lobe (MTL) plays a critical role in working memory (WM) and perception, but these results have been highly controversial given the traditional association of MTL with long-term memory. We review the research and highlight important factors that need to be considered in determining the role of MTL in WM including set-size of used stimuli and feature complexity and/or feature conjunctions/bindings embedded in those stimuli. These factors relate to hierarchical and, accordingly, domain-specific theories of functional organization within the temporal lobe. In addition, one must consider process-specific theories too, because two key processes commonly understood to contribute recognition memory, namely, recollection and familiarity, also have robust support from neurophysiological and neuroimaging research as to their functional dissociations within MTL. PFC has long been heavily implicated in WM; however, relatively less is known about how the PFC contributes to recollection and familiarity, although dynamic prefrontal coding models in WM may help to explain their neural mechanisms. The MTL and PFC are heavily interconnected and do not operate independently in underlying WM. We propose that investigation of the interactions between these two regions in WM, particularly their coordinated neural activities, and the modeling of such interactions, will be crucial for the advancing understanding of the neural mechanisms of WM.

Damage to the human dentate gyrus impairs the perceptual discrimination of complex, novel objects

The hippocampus (HPC), and the dentate gyrus (DG) subregion in particular, is purported to be a pattern separator, orthogonally representing similar information so that distinct memories may be formed. The HPC may also be involved in complex perceptual discrimination. It is unclear if this role is limited to spatial/scene stimuli or extends to the discrimination of objects. Also unclear is whether the DG itself contributes to pattern separation beyond memory. BL, an individual with bilateral DG lesions, was previously shown to have poor discrimination of similar, everyday objects in memory. Here, we demonstrate that BL's deficit extends to complex perceptual discrimination of novel objects. Specifically, BL was presented with closely matched possible and impossible objects, which give rise to fundamentally different 3D perceptual representations despite being visually similar. BL performed significantly worse than controls when asked to select an odd object (e.g., impossible) amongst three identical counterpart objects (e.g., possible) presented at different rotations. His deficit was also evident in an atypical eye fixation pattern during this task. In contrast, BL's performance was indistinguishable from that of controls on other tasks involving the same objects, indicating that he could visually differentiate the object pairs, that he perceived the objects holistically in 3D, and that he has only a mild weakness in categorizing object possibility. Furthermore, his performance on standardized neuropsychological measures indicated intact mental rotation, visual-spatial attention, and working memory (visual and auditory). Collectively, these results provide evidence that the DG is necessary for complex perceptual discrimination of novel objects, indicating that the DG might function as a generic pattern separator of a wide range of stimuli within high-level perception, and that its role is not limited to memory.

Texture-like representation of objects in human visual cortex

Humans are exquisitely sensitive to the spatial arrangement of visual features in objects and scenes, but not in visual textures. Category-selective regions in the visual cortex are widely believed to underlie object perception, suggesting such regions should distinguish natural images of objects from synthesized images containing similar visual features in scrambled arrangements. Contrarily, we demonstrate that representations in category-selective cortex do not discriminate natural images from feature-matched scrambles but can discriminate images of different categories, suggesting a texture-like encoding. We find similar insensitivity to feature arrangement in Imagenet-trained deep convolutional neural networks. This suggests the need to reconceptualize the role of category-selective cortex as representing a basis set of complex texture-like features, useful for a myriad of behaviors.

The human visual ability to recognize objects and scenes is widely thought to rely on representations in category-selective regions of the visual cortex. These representations could support object vision by specifically representing objects, or, more simply, by representing complex visual features regardless of the particular spatial arrangement needed to constitute real-world objects, that is, by representing visual textures. To discriminate between these hypotheses, we leveraged an image synthesis approach that, unlike previous methods, provides independent control over the complexity and spatial arrangement of visual features. We found that human observers could easily detect a natural object among synthetic images with similar complex features that were spatially scrambled. However, observer models built from BOLD responses from category-selective regions, as well as a model of macaque inferotemporal cortex and Imagenet-trained deep convolutional neural networks, were all unable to identify the real object. This inability was not due to a lack of signal to noise, as all observer models could predict human performance in image categorization tasks. How then might these texture-like representations in category-selective regions support object perception? An image-specific readout from category-selective cortex yielded a representation that was more selective for natural feature arrangement, showing that the information necessary for natural object discrimination is available. Thus, our results suggest that the role of the human category-selective visual cortex is not to explicitly encode objects but rather to provide a basis set of texture-like features that can be infinitely reconfigured to flexibly learn and identify new object categories.

Activity in perirhinal and entorhinal cortex predicts perceived visual similarities among category exemplars with highest precision

Vision neuroscience has made great strides in understanding the hierarchical organization of object representations along the ventral visual stream (VVS). How VVS representations capture fine-grained visual similarities between objects that observers subjectively perceive has received limited examination so far. In the current study, we addressed this question by focussing on perceived visual similarities among subordinate exemplars of real-world categories. We hypothesized that these perceived similarities are reflected with highest fidelity in neural activity patterns downstream from inferotemporal regions, namely in perirhinal (PrC) and anterolateral entorhinal cortex (alErC) in the medial temporal lobe. To address this issue with functional magnetic resonance imaging (fMRI), we administered a modified 1-back task that required discrimination between category exemplars as well as categorization. Further, we obtained observer-specific ratings of perceived visual similarities, which predicted behavioural discrimination performance during scanning. As anticipated, we found that activity patterns in PrC and alErC predicted the structure of perceived visual similarity relationships among category exemplars, including its observer-specific component, with higher precision than any other VVS region. Our findings provide new evidence that subjective aspects of object perception that rely on fine-grained visual differentiation are reflected with highest fidelity in the medial temporal lobe.

Perception and memory in the medial temporal lobe: Deep learning offers a new lens on an old debate

In this issue of Neuron, Bonnen et al. (2021) use artificial neural networks to resolve a long-standing controversy surrounding the neurocognitive dichotomy between memory and perception. They show that the perirhinal cortex supports performance on tasks that cannot be solved by the ventral visual stream.