Feature-reweighted representational similarity analysis: A method for improving the fit between computational models, brains, and behavior

The Dimensions of dimensionality

Cognitive scientists often infer multidimensional representations from data. Whether the data involve text, neuroimaging, neural networks, or human judgments, researchers frequently infer and analyze latent representational spaces (i.e., embeddings). However, the properties of a latent representation (e.g., prediction performance, interpretability, compactness) depend on the inference procedure, which can vary widely across endeavors. For example, dimensions are not always globally interpretable and the dimensionality of different embeddings may not be readily comparable. Moreover, the dichotomy between multidimensional spaces and purportedly richer representational formats, such as graph representations, is misleading. We review what the different notions of dimension in cognitive science imply for how these latent representations should be used and interpreted.

Fine-grained knowledge about manipulable objects is well-predicted by contrastive language image pre-training

Object recognition is an important ability that relies on distinguishing between similar objects (e.g., deciding which utensil(s) to use at different stages of meal preparation). Recent work describes the fine-grained organization of knowledge about manipulable objects via the study of the constituent dimensions that are most relevant to human behavior, for example, vision, manipulation, and function-based properties. A logical extension of this work concerns whether or not these dimensions are uniquely human, or can be approximated by deep learning. Here, we show that behavioral dimensions are generally well-predicted by CLIP-ViT - a multimodal network trained on a large and diverse set of image-text pairs. Moreover, this model outperforms comparison networks pre-trained on smaller, image-only datasets. These results demonstrate the impressive capacity of CLIP-ViT to approximate fine-grained object knowledge. We discuss the possible sources of this benefit relative to other models (e.g., multimodal vs. image-only pre-training, dataset size, architecture).

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.

Modeling naturalistic face processing in humans with deep convolutional neural networks

Deep convolutional neural networks (DCNNs) trained for face identification can rival and even exceed human-level performance. The ways in which the internal face representations in DCNNs relate to human cognitive representations and brain activity are not well understood. Nearly all previous studies focused on static face image processing with rapid display times and ignored the processing of naturalistic, dynamic information. To address this gap, we developed the largest naturalistic dynamic face stimulus set in human neuroimaging research (700+ naturalistic video clips of unfamiliar faces). We used this naturalistic dataset to compare representational geometries estimated from DCNNs, behavioral responses, and brain responses. We found that DCNN representational geometries were consistent across architectures, cognitive representational geometries were consistent across raters in a behavioral arrangement task, and neural representational geometries in face areas were consistent across brains. Representational geometries in late, fully connected DCNN layers, which are optimized for individuation, were much more weakly correlated with cognitive and neural geometries than were geometries in late-intermediate layers. The late-intermediate face-DCNN layers successfully matched cognitive representational geometries, as measured with a behavioral arrangement task that primarily reflected categorical attributes, and correlated with neural representational geometries in known face-selective topographies. Our study suggests that current DCNNs successfully capture neural cognitive processes for categorical attributes of faces but less accurately capture individuation and dynamic features.

Multivariate fMRI responses in superior temporal cortex predict visual contributions to, and individual differences in, the intelligibility of noisy speech

Humans have the unique ability to decode the rapid stream of language elements that constitute speech, even when it is contaminated by noise. Two reliable observations about noisy speech perception are that seeing the face of the talker improves intelligibility and the existence of individual differences in the ability to perceive noisy speech. We introduce a multivariate BOLD fMRI measure that explains both observations. In two independent fMRI studies, clear and noisy speech was presented in visual, auditory and audiovisual formats to thirty-seven participants who rated intelligibility. An event-related design was used to sort noisy speech trials by their intelligibility. Individual-differences multidimensional scaling was applied to fMRI response patterns in superior temporal cortex and the dissimilarity between responses to clear speech and noisy (but intelligible) speech was measured. Neural dissimilarity was less for audiovisual speech than auditory-only speech, corresponding to the greater intelligibility of noisy audiovisual speech. Dissimilarity was less in participants with better noisy speech perception, corresponding to individual differences. These relationships held for both single word and entire sentence stimuli, suggesting that they were driven by intelligibility rather than the specific stimuli tested. A neural measure of perceptual intelligibility may aid in the development of strategies for helping those with impaired speech perception.

Functionally analogous body- and animacy-responsive areas are present in the dog (Canis familiaris) and human occipito-temporal lobe

Comparing the neural correlates of socio-cognitive skills across species provides insights into the evolution of the social brain and has revealed face- and body-sensitive regions in the primate temporal lobe. Although from a different lineage, dogs share convergent visuo-cognitive skills with humans and a temporal lobe which evolved independently in carnivorans. We investigated the neural correlates of face and body perception in dogs (N = 15) and humans (N = 40) using functional MRI. Combining univariate and multivariate analysis approaches, we found functionally analogous occipito-temporal regions involved in the perception of animate entities and bodies in both species and face-sensitive regions in humans. Though unpredicted, we also observed neural representations of faces compared to inanimate objects, and dog compared to human bodies in dog olfactory regions. These findings shed light on the evolutionary foundations of human and dog social cognition and the predominant role of the temporal lobe.

The neuroconnectionist research programme

Artificial neural networks (ANNs) inspired by biology are beginning to be widely used to model behavioural and neural data, an approach we call 'neuroconnectionism'. ANNs have been not only lauded as the current best models of information processing in the brain but also criticized for failing to account for basic cognitive functions. In this Perspective article, we propose that arguing about the successes and failures of a restricted set of current ANNs is the wrong approach to assess the promise of neuroconnectionism for brain science. Instead, we take inspiration from the philosophy of science, and in particular from Lakatos, who showed that the core of a scientific research programme is often not directly falsifiable but should be assessed by its capacity to generate novel insights. Following this view, we present neuroconnectionism as a general research programme centred around ANNs as a computational language for expressing falsifiable theories about brain computation. We describe the core of the programme, the underlying computational framework and its tools for testing specific neuroscientific hypotheses and deriving novel understanding. Taking a longitudinal view, we review past and present neuroconnectionist projects and their responses to challenges and argue that the research programme is highly progressive, generating new and otherwise unreachable insights into the workings of the brain.

Every individual makes a difference: A trinity derived from linking individual brain morphometry, connectivity and mentalising ability

Mentalising ability, indexed as the ability to understand others' beliefs, feelings, intentions, thoughts and traits, is a pivotal and fundamental component of human social cognition. However, considering the multifaceted nature of mentalising ability, little research has focused on characterising individual differences in different mentalising components. And even less research has been devoted to investigating how the variance in the structural and functional patterns of the amygdala and hippocampus, two vital subcortical regions of the "social brain", are related to inter-individual variability in mentalising ability. Here, as a first step toward filling these gaps, we exploited inter-subject representational similarity analysis (IS-RSA) to assess relationships between amygdala and hippocampal morphometry (surface-based multivariate morphometry statistics, MMS), connectivity (resting-state functional connectivity, rs-FC) and mentalising ability (interactive mentalisation questionnaire [IMQ] scores) across the participants ( N = 24 $$ N=24 $$ ). In IS-RSA, we proposed a novel pipeline, that is, computing patching and pooling operations-based surface distance (CPP-SD), to obtain a decent representation for high-dimensional MMS data. On this basis, we found significant correlations (i.e., second-order isomorphisms) between these three distinct modalities, indicating that a trinity existed in idiosyncratic patterns of brain morphometry, connectivity and mentalising ability. Notably, a region-related mentalising specificity emerged from these associations: self-self and self-other mentalisation are more related to the hippocampus, while other-self mentalisation shows a closer link with the amygdala. Furthermore, by utilising the dyadic regression analysis, we observed significant interactions such that subject pairs with similar morphometry had even greater mentalising similarity if they were also similar in rs-FC. Altogether, we demonstrated the feasibility and illustrated the promise of using IS-RSA to study individual differences, deepening our understanding of how individual brains give rise to their mentalising abilities.

Deep Neural Networks and Visuo-Semantic Models Explain Complementary Components of Human Ventral-Stream Representational Dynamics

Deep neural networks (DNNs) are promising models of the cortical computations supporting human object recognition. However, despite their ability to explain a significant portion of variance in neural data, the agreement between models and brain representational dynamics is far from perfect. We address this issue by asking which representational features are currently unaccounted for in neural time series data, estimated for multiple areas of the ventral stream via source-reconstructed magnetoencephalography data acquired in human participants (nine females, six males) during object viewing. We focus on the ability of visuo-semantic models, consisting of human-generated labels of object features and categories, to explain variance beyond the explanatory power of DNNs alone. We report a gradual reversal in the relative importance of DNN versus visuo-semantic features as ventral-stream object representations unfold over space and time. Although lower-level visual areas are better explained by DNN features starting early in time (at 66 ms after stimulus onset), higher-level cortical dynamics are best accounted for by visuo-semantic features starting later in time (at 146 ms after stimulus onset). Among the visuo-semantic features, object parts and basic categories drive the advantage over DNNs. These results show that a significant component of the variance unexplained by DNNs in higher-level cortical dynamics is structured and can be explained by readily nameable aspects of the objects. We conclude that current DNNs fail to fully capture dynamic representations in higher-level human visual cortex and suggest a path toward more accurate models of ventral-stream computations.SIGNIFICANCE STATEMENT When we view objects such as faces and cars in our visual environment, their neural representations dynamically unfold over time at a millisecond scale. These dynamics reflect the cortical computations that support fast and robust object recognition. DNNs have emerged as a promising framework for modeling these computations but cannot yet fully account for the neural dynamics. Using magnetoencephalography data acquired in human observers during object viewing, we show that readily nameable aspects of objects, such as 'eye', 'wheel', and 'face', can account for variance in the neural dynamics over and above DNNs. These findings suggest that DNNs and humans may in part rely on different object features for visual recognition and provide guidelines for model improvement.

A unifying model for discordant and concordant results in human neuroimaging studies of facial viewpoint selectivity

Our ability to recognize faces regardless of viewpoint is a key property of the primate visual system. Traditional theories hold that facial viewpoint is represented by view-selective mechanisms at early visual processing stages and that representations become increasingly tolerant to viewpoint changes in higher-level visual areas. Newer theories, based on single-neuron monkey electrophysiological recordings, suggest an additional intermediate processing stage invariant to mirror-symmetric face views. Consistent with traditional theories, human studies combining neuroimaging and multivariate pattern analysis (MVPA) methods have provided evidence of view-selectivity in early visual cortex. However, contradictory results have been reported in higher-level visual areas concerning the existence in humans of mirror-symmetrically tuned representations. We believe these results reflect low-level stimulus confounds and data analysis choices. To probe for low-level confounds, we analyzed images from two popular face databases. Analyses of mean image luminance and contrast revealed biases across face views described by even polynomials-i.e., mirror-symmetric. To explain major trends across human neuroimaging studies of viewpoint selectivity, we constructed a network model that incorporates three biological constraints: cortical magnification, convergent feedforward projections, and interhemispheric connections. Given the identified low-level biases, we show that a gradual increase of interhemispheric connections across network layers is sufficient to replicate findings of mirror-symmetry in high-level processing stages, as well as view-tuning in early processing stages. Data analysis decisions-pattern dissimilarity measure and data recentering-accounted for the variable observation of mirror-symmetry in late processing stages. The model provides a unifying explanation of MVPA studies of viewpoint selectivity. We also show how common analysis choices can lead to erroneous conclusions.

Exploring the role of structuralist methodology in the neuroscience of consciousness: a defense and analysis

Traditional contrastive analysis has been the foundation of consciousness science, but its limitations due to the lack of a reliable method for measuring states of consciousness have prompted the exploration of alternative approaches. Structuralist theories have gained attention as an alternative that focuses on the structural properties of phenomenal experience and seeks to identify their neural encoding via structural similarities between quality spaces and neural state spaces. However, the intertwining of philosophical assumptions about structuralism and structuralist methodology may pose a challenge to those who are skeptical of the former. In this paper, I offer an analysis and defense of structuralism as a methodological approach in consciousness science, which is partly independent of structuralist assumptions on the nature of consciousness. By doing so, I aim to make structuralist methodology more accessible to a broader scientific and philosophical audience. I situate methodological structuralism in the context of questions concerning mental representation, psychophysical measurement, holism, and functional relevance of neural processes. At last, I analyze the relationship between the structural approach and the distinction between conscious and unconscious states.

Improving the accuracy of single-trial fMRI response estimates using GLMsingle

Advances in artificial intelligence have inspired a paradigm shift in human neuroscience, yielding large-scale functional magnetic resonance imaging (fMRI) datasets that provide high-resolution brain responses to thousands of naturalistic visual stimuli. Because such experiments necessarily involve brief stimulus durations and few repetitions of each stimulus, achieving sufficient signal-to-noise ratio can be a major challenge. We address this challenge by introducing GLMsingle, a scalable, user-friendly toolbox available in MATLAB and Python that enables accurate estimation of single-trial fMRI responses (glmsingle.org). Requiring only fMRI time-series data and a design matrix as inputs, GLMsingle integrates three techniques for improving the accuracy of trial-wise general linear model (GLM) beta estimates. First, for each voxel, a custom hemodynamic response function (HRF) is identified from a library of candidate functions. Second, cross-validation is used to derive a set of noise regressors from voxels unrelated to the experiment. Third, to improve the stability of beta estimates for closely spaced trials, betas are regularized on a voxel-wise basis using ridge regression. Applying GLMsingle to the Natural Scenes Dataset and BOLD5000, we find that GLMsingle substantially improves the reliability of beta estimates across visually-responsive cortex in all subjects. Comparable improvements in reliability are also observed in a smaller-scale auditory dataset from the StudyForrest experiment. These improvements translate into tangible benefits for higher-level analyses relevant to systems and cognitive neuroscience. We demonstrate that GLMsingle: (i) helps decorrelate response estimates between trials nearby in time; (ii) enhances representational similarity between subjects within and across datasets; and (iii) boosts one-versus-many decoding of visual stimuli. GLMsingle is a publicly available tool that can significantly improve the quality of past, present, and future neuroimaging datasets sampling brain activity across many experimental conditions.