How does the brain solve visual object recognition?

Deeper neural network models better reflect how humans cope with contrast variation in object recognition

Visual inputs are far from ideal in everyday situations such as in the fog where the contrasts of input stimuli are low. However, human perception remains relatively robust to contrast variations. To provide insights about the underlying mechanisms of contrast invariance, we addressed two questions. Do contrast effects disappear along the visual hierarchy? Do later stages of the visual hierarchy contribute to contrast invariance? We ran a behavioral experiment where we manipulated the level of stimulus contrast and the involvement of higher-level visual areas through immediate and delayed backward masking of the stimulus. Backward masking led to significant drop in performance in our visual categorization task, supporting the role of higher-level visual areas in contrast invariance. To obtain mechanistic insights, we ran the same categorization task on three state-of the-art computational models of human vision each with a different depth in visual hierarchy. We found contrast effects all along the visual hierarchy, no matter how far into the hierarchy. Moreover, that final layers of deeper hierarchical models, which had been shown to be best models of final stages of the visual system, coped with contrast effects more effectively. These results suggest that, while contrast effects reach the final stages of the hierarchy, those stages play a significant role in compensating for contrast variations in the visual system.

Understanding Human Object Vision: A Picture Is Worth a Thousand Representations

Objects are the core meaningful elements in our visual environment. Classic theories of object vision focus upon object recognition and are elegant and simple. Some of their proposals still stand, yet the simplicity is gone. Recent evolutions in behavioral paradigms, neuroscientific methods, and computational modeling have allowed vision scientists to uncover the complexity of the multidimensional representational space that underlies object vision. We review these findings and propose that the key to understanding this complexity is to relate object vision to the full repertoire of behavioral goals that underlie human behavior, running far beyond object recognition. There might be no such thing as core object recognition, and if it exists, then its importance is more limited than traditionally thought.

Role of line junctions in expert object recognition: The case of musical notation

Line junctions are well-known to be important for real-world object recognition, and sensitivity to line junctions is enhanced with perceptual experience with an object category. However, it remains unclear whether these very simple visual features are involved in expert object representations at the neural level, and if yes, at what level(s) they are involved. In this EEG study, 31 music reading experts and 31 novices performed a one-back task with intact musical notation, musical notation with line junctions removed and pseudo-letters. We observed more separable neural representations of musical notation from pseudo-letter for experts than for novices when line junctions were present and during 180-280 ms after stimulus onset. Also, the presence of line junctions was better decoded in experts than in novices during 320-580 ms, and the decoding accuracy in this time window predicted the behavioral recognition advantage of musical notation when line junctions were present. These suggest that, with perceptual expertise, line junctions are more involved in category selective representation of objects, and are more explicitly represented in later stages of processing to support expert recognition performance.

Altered Functional Connectivity Density in Type 2 Diabetes Mellitus with and without Mild Cognitive Impairment

Although disturbed functional connectivity is known to be a factor influencing cognitive impairment, the neuropathological mechanisms underlying the cognitive impairment caused by type 2 diabetes mellitus (T2DM) remain unclear. To characterize the neural mechanisms underlying T2DM-related brain damage, we explored the altered functional architecture patterns in different cognitive states in T2DM patients. Thirty-seven T2DM patients with normal cognitive function (DMCN), 40 T2DM patients with mild cognitive impairment (MCI) (DMCI), and 40 healthy controls underwent neuropsychological assessments and resting-state functional MRI examinations. Functional connectivity density (FCD) analysis was performed, and the relationship between abnormal FCD and clinical/cognitive variables was assessed. The regions showing abnormal FCD in T2DM patients were mainly located in the temporal lobe and cerebellum, but the abnormal functional architecture was more extensive in DMCI patients. Moreover, in comparison with the DMCN group, DMCI patients showed reduced long-range FCD in the left superior temporal gyrus (STG), which was correlated with the Rey auditory verbal learning test score in all T2DM patients. Thus, DMCI patients show functional architecture abnormalities in more brain regions involved in higher-level cognitive function (executive function and auditory memory function), and the left STG may be involved in the neuropathology of auditory memory in T2DM patients. These findings provide some new insights into understanding the neural mechanisms underlying T2DM-related cognitive impairment.

Transformation of acoustic information to sensory decision variables in the parietal cortex

The process by which sensory evidence contributes to perceptual choices requires an understanding of its transformation into decision variables. Here, we address this issue by evaluating the neural representation of acoustic information in the auditory cortex-recipient parietal cortex, while gerbils either performed a two-alternative forced-choice auditory discrimination task or while they passively listened to identical acoustic stimuli. During task engagement, stimulus identity decoding performance from simultaneously recorded parietal neurons significantly correlated with psychometric sensitivity. In contrast, decoding performance during passive listening was significantly reduced. Principal component and geometric analyses revealed the emergence of low-dimensional encoding of linearly separable manifolds with respect to stimulus identity and decision, but only during task engagement. These findings confirm that the parietal cortex mediates a transition of acoustic representations into decision-related variables. Finally, using a clustering analysis, we identified three functionally distinct subpopulations of neurons that each encoded task-relevant information during separate temporal segments of a trial. Taken together, our findings demonstrate how parietal cortex neurons integrate and transform encoded auditory information to guide sound-driven perceptual decisions.

Feature Map Augmentation to Improve Scale Invariance in Convolutional Neural Networks

Introducing variation in the training dataset through data augmentation has been a popular technique to make Convolutional Neural Networks (CNNs) spatially invariant but leads to increased dataset volume and computation cost. Instead of data augmentation, augmentation of feature maps is proposed to introduce variations in the features extracted by a CNN. To achieve this, a rotation transformer layer called Rotation Invariance Transformer (RiT) is developed, which applies rotation transformation to augment CNN features. The RiT layer can be used to augment output features from any convolution layer within a CNN. However, its maximum effectiveness is shown when placed at the output end of final convolution layer. We test RiT in the application of scale-invariance where we attempt to classify scaled images from benchmark datasets. Our results show promising improvements in the networks ability to be scale invariant whilst keeping the model computation cost low.

Visual category representations in the infant brain

Visual categorization is a human core cognitive capacity1,2 that depends on the development of visual category representations in the infant brain.3,4,5,6,7 However, the exact nature of infant visual category representations and their relationship to the corresponding adult form remains unknown.8 Our results clarify the nature of visual category representations from electroencephalography (EEG) data in 6- to 8-month-old infants and their developmental trajectory toward adult maturity in the key characteristics of temporal dynamics,2,9 representational format,10,11,12 and spectral properties.13,14 Temporal dynamics change from slowly emerging, developing representations in infants to quickly emerging, complex representations in adults. Despite those differences, infants and adults already partly share visual category representations. The format of infants' representations is visual features of low to intermediate complexity, whereas adults' representations also encode high-complexity features. Theta band activity contributes to visual category representations in infants, and these representations are shifted to the alpha/beta band in adults. Together, we reveal the developmental neural basis of visual categorization in humans, show how information transmission channels change in development, and demonstrate the power of advanced multivariate analysis techniques in infant EEG research for theory building in developmental cognitive science.

Personalized visual encoding model construction with small data

Quantifying population heterogeneity in brain stimuli-response mapping may allow insight into variability in bottom-up neural systems that can in turn be related to individual's behavior or pathological state. Encoding models that predict brain responses to stimuli are one way to capture this relationship. However, they generally need a large amount of fMRI data to achieve optimal accuracy. Here, we propose an ensemble approach to create encoding models for novel individuals with relatively little data by modeling each subject's predicted response vector as a linear combination of the other subjects' predicted response vectors. We show that these ensemble encoding models trained with hundreds of image-response pairs, achieve accuracy not different from models trained on 20,000 image-response pairs. Importantly, the ensemble encoding models preserve patterns of inter-individual differences in the image-response relationship. We also show the proposed approach is robust against domain shift by validating on data with a different scanner and experimental setup. Additionally, we show that the ensemble encoding models are able to discover the inter-individual differences in various face areas' responses to images of animal vs human faces using a recently developed NeuroGen framework. Our approach shows the potential to use existing densely-sampled data, i.e. large amounts of data collected from a single individual, to efficiently create accurate, personalized encoding models and, subsequently, personalized optimal synthetic images for new individuals scanned under different experimental conditions.

Visualizing advances in the future of primate neuroscience research

Future neuroscience and biomedical projects involving non-human primates (NHPs) remain essential in our endeavors to understand the complexities and functioning of the mammalian central nervous system. In so doing, the NHP neuroscience researcher must be allowed to incorporate state-of-the-art technologies, including the use of novel viral vectors, gene therapy and transgenic approaches to answer continuing and emerging research questions that can only be addressed in NHP research models. This perspective piece captures these emerging technologies and some specific research questions they can address. At the same time, we highlight some current caveats to global NHP research and collaborations including the lack of common ethical and regulatory frameworks for NHP research, the limitations involving animal transportation and exports, and the ongoing influence of activist groups opposed to NHP research.

Long-range functional loops in the mouse olfactory system and their roles in computing odor identity

Elucidating the neural circuits supporting odor identification remains an open challenge. Here, we analyze the contribution of the two output cell types of the mouse olfactory bulb (mitral and tufted cells) to decode odor identity and concentration and its dependence on top-down feedback from their respective major cortical targets: piriform cortex versus anterior olfactory nucleus. We find that tufted cells substantially outperform mitral cells in decoding both odor identity and intensity. Cortical feedback selectively regulates the activity of its dominant bulb projection cell type and implements different computations. Piriform feedback specifically restructures mitral responses, whereas feedback from the anterior olfactory nucleus preferentially controls the gain of tufted representations without altering their odor tuning. Our results identify distinct functional loops involving the mitral and tufted cells and their cortical targets. We suggest that in addition to the canonical mitral-to-piriform pathway, tufted cells and their target regions are ideally positioned to compute odor identity.

Efficient processing of natural scenes in visual cortex

Neural circuits in the periphery of the visual, auditory, and olfactory systems are believed to use limited resources efficiently to represent sensory information by adapting to the statistical structure of the natural environment. This "efficient coding" principle has been used to explain many aspects of early visual circuits including the distribution of photoreceptors, the mosaic geometry and center-surround structure of retinal receptive fields, the excess OFF pathways relative to ON pathways, saccade statistics, and the structure of simple cell receptive fields in V1. We know less about the extent to which such adaptations may occur in deeper areas of cortex beyond V1. We thus review recent developments showing that the perception of visual textures, which depends on processing in V2 and beyond in mammals, is adapted in rats and humans to the multi-point statistics of luminance in natural scenes. These results suggest that central circuits in the visual brain are adapted for seeing key aspects of natural scenes. We conclude by discussing how adaptation to natural temporal statistics may aid in learning and representing visual objects, and propose two challenges for the future: (1) explaining the distribution of shape sensitivity in the ventral visual stream from the statistics of object shape in natural images, and (2) explaining cell types of the vertebrate retina in terms of feature detectors that are adapted to the spatio-temporal structures of natural stimuli. We also discuss how new methods based on machine learning may complement the normative, principles-based approach to theoretical neuroscience.

Dissociation and hierarchy of human visual pathways for simultaneously coding facial identity and expression

Humans have an extraordinary ability to recognize facial expression and identity from a single face simultaneously and effortlessly, however, the underlying neural computation is not well understood. Here, we optimized a multi-task deep neural network to classify facial expression and identity simultaneously. Under various optimization training strategies, the best-performing model consistently showed 'share-separate' organization. The two separate branches of the best-performing model also exhibited distinct abilities to categorize facial expression and identity, and these abilities increased along the facial expression or identity branches toward high layers. By comparing the representational similarities between the best-performing model and functional magnetic resonance imaging (fMRI) responses in the human visual cortex to the same face stimuli, the face-selective posterior superior temporal sulcus (pSTS) in the dorsal visual cortex was significantly correlated with layers in the expression branch of the model, and the anterior inferotemporal cortex (aIT) and anterior fusiform face area (aFFA) in the ventral visual cortex were significantly correlated with layers in the identity branch of the model. Besides, the aFFA and aIT better matched the high layers of the model, while the posterior FFA (pFFA) and occipital facial area (OFA) better matched the middle and early layers of the model, respectively. Overall, our study provides a task-optimization computational model to better understand the neural mechanism underlying face recognition, which suggest that similar to the best-performing model, the human visual system exhibits both dissociated and hierarchical neuroanatomical organization when simultaneously coding facial identity and expression.

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.

Learning dynamics of deep linear networks with multiple pathways

Not only have deep networks become standard in machine learning, they are increasingly of interest in neuroscience as models of cortical computation that capture relationships between structural and functional properties. In addition they are a useful target of theoretical research into the properties of network computation. Deep networks typically have a serial or approximately serial organization across layers, and this is often mirrored in models that purport to represent computation in mammalian brains. There are, however, multiple examples of parallel pathways in mammalian brains. In some cases, such as the mouse, the entire visual system appears arranged in a largely parallel, rather than serial fashion. While these pathways may be formed by differing cost functions that drive different computations, here we present a new mathematical analysis of learning dynamics in networks that have parallel computational pathways driven by the same cost function. We use the approximation of deep linear networks with large hidden layer sizes to show that, as the depth of the parallel pathways increases, different features of the training set (defined by the singular values of the input-output correlation) will typically concentrate in one of the pathways. This result is derived analytically and demonstrated with numerical simulation with both linear and non-linear networks. Thus, rather than sharing stimulus and task features across multiple pathways, parallel network architectures learn to produce sharply diversified representations with specialized and specific pathways, a mechanism which may hold important consequences for codes in both biological and artificial systems.

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.

Magnetoencephalography recordings reveal the spatiotemporal dynamics of recognition memory for complex versus simple auditory sequences

Auditory recognition is a crucial cognitive process that relies on the organization of single elements over time. However, little is known about the spatiotemporal dynamics underlying the conscious recognition of auditory sequences varying in complexity. To study this, we asked 71 participants to learn and recognize simple tonal musical sequences and matched complex atonal sequences while their brain activity was recorded using magnetoencephalography (MEG). Results reveal qualitative changes in neural activity dependent on stimulus complexity: recognition of tonal sequences engages hippocampal and cingulate areas, whereas recognition of atonal sequences mainly activates the auditory processing network. Our findings reveal the involvement of a cortico-subcortical brain network for auditory recognition and support the idea that stimulus complexity qualitatively alters the neural pathways of recognition memory.

Zoom-In Neural Network Deep-Learning Model for Alzheimer's Disease Assessments

Deep neural networks have been successfully applied to generate predictive patterns from medical and diagnostic data. This paper presents an approach for assessing persons with Alzheimer's disease (AD) mild cognitive impairment (MCI), compared with normal control (NC) persons, using the zoom-in neural network (ZNN) deep-learning algorithm. ZNN stacks a set of zoom-in learning units (ZLUs) in a feedforward hierarchy without backpropagation. The resting-state fMRI (rs-fMRI) dataset for AD assessments was obtained from the Alzheimer's Disease Neuroimaging Initiative (ADNI). The Automated Anatomical Labeling (AAL-90) atlas, which provides 90 neuroanatomical functional regions, was used to assess and detect the implicated regions in the course of AD. The features of the ZNN are extracted from the 140-time series rs-fMRI voxel values in a region of the brain. ZNN yields the three classification accuracies of AD versus MCI and NC, NC versus AD and MCI, and MCI versus AD and NC of 97.7%, 84.8%, and 72.7%, respectively, with the seven discriminative regions of interest (ROIs) in the AAL-90.

Invariance of object detection in untrained deep neural networks

The ability to perceive visual objects with various types of transformations, such as rotation, translation, and scaling, is crucial for consistent object recognition. In machine learning, invariant object detection for a network is often implemented by augmentation with a massive number of training images, but the mechanism of invariant object detection in biological brains—how invariance arises initially and whether it requires visual experience—remains elusive. Here, using a model neural network of the hierarchical visual pathway of the brain, we show that invariance of object detection can emerge spontaneously in the complete absence of learning. First, we found that units selective to a particular object class arise in randomly initialized networks even before visual training. Intriguingly, these units show robust tuning to images of each object class under a wide range of image transformation types, such as viewpoint rotation. We confirmed that this “innate” invariance of object selectivity enables untrained networks to perform an object-detection task robustly, even with images that have been significantly modulated. Our computational model predicts that invariant object tuning originates from combinations of non-invariant units via random feedforward projections, and we confirmed that the predicted profile of feedforward projections is observed in untrained networks. Our results suggest that invariance of object detection is an innate characteristic that can emerge spontaneously in random feedforward networks.

Predicting Identity-Preserving Object Transformations in Human Posterior Parietal Cortex and Convolutional Neural Networks

Previous research shows that, within human occipito-temporal cortex (OTC), we can use a general linear mapping function to link visual object responses across nonidentity feature changes, including Euclidean features (e.g., position and size) and non-Euclidean features (e.g., image statistics and spatial frequency). Although the learned mapping is capable of predicting responses of objects not included in training, these predictions are better for categories included than those not included in training. These findings demonstrate a near-orthogonal representation of object identity and nonidentity features throughout human OTC. Here, we extended these findings to examine the mapping across both Euclidean and non-Euclidean feature changes in human posterior parietal cortex (PPC), including functionally defined regions in inferior and superior intraparietal sulcus. We additionally examined responses in five convolutional neural networks (CNNs) pretrained with object classification, as CNNs are considered as the current best model of the primate ventral visual system. We separately compared results from PPC and CNNs with those of OTC. We found that a linear mapping function could successfully link object responses in different states of nonidentity transformations in human PPC and CNNs for both Euclidean and non-Euclidean features. Overall, we found that object identity and nonidentity features are represented in a near-orthogonal, rather than complete-orthogonal, manner in PPC and CNNs, just like they do in OTC. Meanwhile, some differences existed among OTC, PPC, and CNNs. These results demonstrate the similarities and differences in how visual object information across an identity-preserving image transformation may be represented in OTC, PPC, and CNNs.

Understanding transformation tolerant visual object representations in the human brain and convolutional neural networks

Forming transformation-tolerant object representations is critical to high-level primate vision. Despite its significance, many details of tolerance in the human brain remain unknown. Likewise, despite the ability of convolutional neural networks (CNNs) to exhibit human-like object categorization performance, whether CNNs form tolerance similar to that of the human brain is unknown. Here we provide the first comprehensive documentation and comparison of three tolerance measures in the human brain and CNNs. We measured fMRI responses from human ventral visual areas to real-world objects across both Euclidean and non-Euclidean feature changes. In single fMRI voxels in higher visual areas, we observed robust object response rank-order preservation across feature changes. This is indicative of functional smoothness in tolerance at the fMRI meso-scale level that has never been reported before. At the voxel population level, we found highly consistent object representational structure across feature changes towards the end of ventral processing. Rank-order preservation, consistency, and a third tolerance measure, cross-decoding success (i.e., a linear classifier's ability to generalize performance across feature changes) showed an overall tight coupling. These tolerance measures were in general lower for Euclidean than non-Euclidean feature changes in lower visual areas, but increased over the course of ventral processing for all feature changes. These characteristics of tolerance, however, were absent in eight CNNs pretrained with ImageNet images with varying network architecture, depth, the presence/absence of recurrent processing, or whether a network was pretrained with the original or stylized ImageNet images that encouraged shape processing. CNNs do not appear to develop the same kind of tolerance as the human brain over the course of visual processing.

Under-resourced or overloaded? Rethinking working memory deficits in developmental language disorder

Dominant theoretical accounts of developmental language disorder (DLD) commonly invoke working memory capacity limitations. In the current report, we present an alternative view: That working memory in DLD is not under-resourced but overloaded due to operating on speech representations with low discriminability. This account is developed through computational simulations involving deep convolutional neural networks trained on spoken word spectrograms in which information is either retained to mimic typical development or degraded to mimic the auditory processing deficits identified among some children with DLD. We assess not only spoken word recognition accuracy and predictive probability and entropy (i.e., predictive distribution spread), but also use mean-field-theory based manifold analysis to assess; (a) internal speech representation dimensionality and (b) classification capacity, a measure of the networks' ability to isolate any given internal speech representation that is used as a proxy for attentional control. We show that instantiating a low-level auditory processing deficit results in the formation of internal speech representations with atypically high dimensionality, and that classification capacity is exhausted due to low representation separability. These representation and control deficits underpin not only lower performance accuracy but also greater uncertainty even when making accurate predictions in a simulated spoken word recognition task (i.e., predictive distributions with low maximum probability and high entropy), which replicates the response delays and word finding difficulties often seen in DLD. Overall, these simulations demonstrate a theoretical account of speech representation and processing deficits in DLD in which working memory capacity limitations play no causal role. (PsycInfo Database Record (c) 2023 APA, all rights reserved).

Neural representational geometry underlies few-shot concept learning

Understanding the neural basis of the remarkable human cognitive capacity to learn novel concepts from just one or a few sensory experiences constitutes a fundamental problem. We propose a simple, biologically plausible, mathematically tractable, and computationally powerful neural mechanism for few-shot learning of naturalistic concepts. We posit that the concepts that can be learned from few examples are defined by tightly circumscribed manifolds in the neural firing-rate space of higher-order sensory areas. We further posit that a single plastic downstream readout neuron learns to discriminate new concepts based on few examples using a simple plasticity rule. We demonstrate the computational power of our proposal by showing that it can achieve high few-shot learning accuracy on natural visual concepts using both macaque inferotemporal cortex representations and deep neural network (DNN) models of these representations and can even learn novel visual concepts specified only through linguistic descriptors. Moreover, we develop a mathematical theory of few-shot learning that links neurophysiology to predictions about behavioral outcomes by delineating several fundamental and measurable geometric properties of neural representations that can accurately predict the few-shot learning performance of naturalistic concepts across all our numerical simulations. This theory reveals, for instance, that high-dimensional manifolds enhance the ability to learn new concepts from few examples. Intriguingly, we observe striking mismatches between the geometry of manifolds in the primate visual pathway and in trained DNNs. We discuss testable predictions of our theory for psychophysics and neurophysiological experiments.

Micro-scale functional modules in the human temporal lobe

The sensory cortices of many mammals are often organized into modules in the form of cortical columns, yet whether modular organization at this spatial scale is a general property of the human neocortex is unknown. The strongest evidence for modularity arises when measures of connectivity, structure, and function converge. Here we use microelectrode recordings in humans to examine functional connectivity and neuronal spiking responses in order to assess modularity in submillimeter scale networks. We find that the human temporal lobe consists of temporally persistent spatially compact modules approximately 1.3mm in diameter. Functionally, the information coded by single neurons during an image categorization task is more similar for neurons belonging to the same module than for neurons from different modules. The geometry, connectivity, and spiking responses of these local cortical networks provide converging evidence that the human temporal lobe is organized into functional modules at the micro scale.

Data-driven emergence of convolutional structure in neural networks

Exploiting data invariances is crucial for efficient learning in both artificial and biological neural circuits. Understanding how neural networks can discover appropriate representations capable of harnessing the underlying symmetries of their inputs is thus crucial in machine learning and neuroscience. Convolutional neural networks, for example, were designed to exploit translation symmetry, and their capabilities triggered the first wave of deep learning successes. However, learning convolutions directly from translation-invariant data with a fully connected network has so far proven elusive. Here we show how initially fully connected neural networks solving a discrimination task can learn a convolutional structure directly from their inputs, resulting in localized, space-tiling receptive fields. These receptive fields match the filters of a convolutional network trained on the same task. By carefully designing data models for the visual scene, we show that the emergence of this pattern is triggered by the non-Gaussian, higher-order local structure of the inputs, which has long been recognized as the hallmark of natural images. We provide an analytical and numerical characterization of the pattern formation mechanism responsible for this phenomenon in a simple model and find an unexpected link between receptive field formation and tensor decomposition of higher-order input correlations. These results provide a perspective on the development of low-level feature detectors in various sensory modalities and pave the way for studying the impact of higher-order statistics on learning in neural networks.

Neural Decoding of Visual Information Across Different Neural Recording Modalities and Approaches

Vision plays a peculiar role in intelligence. Visual information, forming a large part of the sensory information, is fed into the human brain to formulate various types of cognition and behaviours that make humans become intelligent agents. Recent advances have led to the development of brain-inspired algorithms and models for machine vision. One of the key components of these methods is the utilization of the computational principles underlying biological neurons. Additionally, advanced experimental neuroscience techniques have generated different types of neural signals that carry essential visual information. Thus, there is a high demand for mapping out functional models for reading out visual information from neural signals. Here, we briefly review recent progress on this issue with a focus on how machine learning techniques can help in the development of models for contending various types of neural signals, from fine-scale neural spikes and single-cell calcium imaging to coarse-scale electroencephalography (EEG) and functional magnetic resonance imaging recordings of brain signals.

Implementasi Deteksi Tepi menggunakan Metode Quadrant Tree Classifier pada Pemisahan Objek Berbasis Digital Image Processing (Studi Kasus Objek Bendera Negara)

Deteksi tepi adalah segmentasi input citra yang bertujuan untuk menentukan tepi dengan menandai bagian detail dari suatu citra. Dari beberapa penelitian sebelumnya belum menunjukkan hasil deteksi untuk dapat memisahkan objek dari pusat citra masukan citra itu sendiri. Tujuan dari penelitian ini adalah melakukan fungsi deteksi tepi dengan membagi menjadi node menggunakan konsep metode Quadran Tree Classifier untuk diterapkan pada studi kasus objek citra berwarna menggunakan bendera negara. Beberapa gambar masukan memiliki tingkat kerumitan dan piksel yang berbeda, antara lain bendera Korea, bendera Wales, dan bendera Indonesia yang berkibar. Metode yang digunakan adalah adopsi struktur data pohon, dimana masing-masing memiliki 4 node dengan jumlah child node yang sama. Jika node memiliki anak, jumlah node harus 4, secara rekursif melakukan loop. Konsep kerja dari metode split and merge segmentation ini. Hasil segmentasi objek digabungkan sesuai dengan homogenitas warna, terutama yang memiliki kerancuan. Penelitian ini menunjukkan bahwa mampu mengamati pemindaian piksel pada citra bendera Korea dan bendera Indonesia yang berkibar, namun level piksel 520 x 347 seperti bendera Wales, metode ini tidak dapat memisahkan antara objek garis yang tidak senggol. Resolusi piksel berpengaruh terhadap total waktu eksekusi segmentasi (menit/detik), total segmentasi yang teridentifikasi dan total warna.

Affordance embeddings for situated language understanding

Much progress in AI over the last decade has been driven by advances in natural language processing technology, in turn facilitated by large datasets and increased computation power used to train large neural language models. These systems demonstrate apparently sophisticated linguistic understanding or generation capabilities, but often fail to transfer their skills to situations they have not encountered before. We argue that computational situated grounding of linguistic information to real or simulated scenarios provide a solution to some of these learning challenges by creating situational representations that both serve as a formal model of the salient phenomena, and contain rich amounts of exploitable, task-appropriate data for training new, flexible computational models. We approach this problem from a neurosymbolic perspective, using multimodal contextual modeling of interactive situations, events, and object properties, particularly afforded behaviors, and habitats, the situations that condition them. These properties are tightly coupled to processes of situated grounding, and herein we discuss we combine neural and symbolic methods with multimodal simulations to create a platform, VoxWorld, for modeling communication in context, and we demonstrate how neural embedding vectors of symbolically-encoded object affordances facilitate transferring knowledge of objects and situations to novel entities, and learning how to recognize and generate linguistic and gestural denotations.

A biologically inspired spiking neural P system in selective visual attention for efficient feature extraction from human motion

Nowadays, human action recognition has become an essential task in health care and other fields. During the last decade, several authors have developed algorithms for human activity detection and recognition by exploiting at the maximum the high-performance computing devices to improve the quality and efficiency of their results. However, in real-time and practical human action recognition applications, the simulation of these algorithms exceed the capacity of current computer systems by considering several factors, such as camera movement, complex scene and occlusion. One potential solution to decrease the computational complexity in the human action detection and recognition can be found in the nature of the human visual perception. Specifically, this process is called selective visual attention. Inspired by this neural phenomena, we propose for the first time a spiking neural P system for efficient feature extraction from human motion. Specifically, we propose this neural structure to carry out a pre-processing stage since many studies have revealed that an analysis of visual information of the human brain proceeds in a sequence of operations, in which each one is applied to a specific location or locations. In this way, this specialized processing have allowed to focus the recognition of the objects in a simpler manner. To create a compact and high speed spiking neural P system, we use their cutting-edge variants, such as rules on the synapses, communication on request and astrocyte-like control. Our results have demonstrated that the use of the proposed neural P system increases significantly the performance of low-computational complexity neural classifiers up to more 97% in the human action recognition.

Guiding visual attention in deep convolutional neural networks based on human eye movements

Deep Convolutional Neural Networks (DCNNs) were originally inspired by principles of biological vision, have evolved into best current computational models of object recognition, and consequently indicate strong architectural and functional parallelism with the ventral visual pathway throughout comparisons with neuroimaging and neural time series data. As recent advances in deep learning seem to decrease this similarity, computational neuroscience is challenged to reverse-engineer the biological plausibility to obtain useful models. While previous studies have shown that biologically inspired architectures are able to amplify the human-likeness of the models, in this study, we investigate a purely data-driven approach. We use human eye tracking data to directly modify training examples and thereby guide the models’ visual attention during object recognition in natural images either toward or away from the focus of human fixations. We compare and validate different manipulation types (i.e., standard, human-like, and non-human-like attention) through GradCAM saliency maps against human participant eye tracking data. Our results demonstrate that the proposed guided focus manipulation works as intended in the negative direction and non-human-like models focus on significantly dissimilar image parts compared to humans. The observed effects were highly category-specific, enhanced by animacy and face presence, developed only after feedforward processing was completed, and indicated a strong influence on face detection. With this approach, however, no significantly increased human-likeness was found. Possible applications of overt visual attention in DCNNs and further implications for theories of face detection are discussed.

Low and high frequency intracranial neural signals match in the human associative cortex

In vivo intracranial recordings of neural activity offer a unique opportunity to understand human brain function. Intracranial electrophysiological (iEEG) activity related to sensory, cognitive or motor events manifests mostly in two types of signals: event-related local field potentials in lower frequency bands (<30 Hz, LF) and broadband activity in the higher end of the frequency spectrum (>30 Hz, High frequency, HF). While most current studies rely exclusively on HF, thought to be more focal and closely related to spiking activity, the relationship between HF and LF signals is unclear, especially in human associative cortex. Here, we provide a large-scale in-depth investigation of the spatial and functional relationship between these 2 signals based on intracranial recordings from 121 individual brains (8000 recording sites). We measure category-selective responses to complex ecologically salient visual stimuli - human faces - across a wide cortical territory in the ventral occipito-temporal cortex (VOTC), with a frequency-tagging method providing high signal-to-noise ratio (SNR) and the same objective quantification of signal and noise for the two frequency ranges. While LF face-selective activity has higher SNR across the VOTC, leading to a larger number of significant electrode contacts especially in the anterior temporal lobe, LF and HF display highly similar spatial, functional, and timing properties. Specifically, and contrary to a widespread assumption, our results point to nearly identical spatial distribution and local spatial extent of LF and HF activity at equal SNR. These observations go a long way towards clarifying the relationship between the two main iEEG signals and reestablish the informative value of LF iEEG to understand human brain function.

Buice M. MouseNet: A biologically constrained convolutional neural network model for the mouse visual cortex

Convolutional neural networks trained on object recognition derive inspiration from the neural architecture of the visual system in mammals, and have been used as models of the feedforward computation performed in the primate ventral stream. In contrast to the deep hierarchical organization of primates, the visual system of the mouse has a shallower arrangement. Since mice and primates are both capable of visually guided behavior, this raises questions about the role of architecture in neural computation. In this work, we introduce a novel framework for building a biologically constrained convolutional neural network model of the mouse visual cortex. The architecture and structural parameters of the network are derived from experimental measurements, specifically the 100-micrometer resolution interareal connectome, the estimates of numbers of neurons in each area and cortical layer, and the statistics of connections between cortical layers. This network is constructed to support detailed task-optimized models of mouse visual cortex, with neural populations that can be compared to specific corresponding populations in the mouse brain. Using a well-studied image classification task as our working example, we demonstrate the computational capability of this mouse-sized network. Given its relatively small size, MouseNet achieves roughly 2/3rds the performance level on ImageNet as VGG16. In combination with the large scale Allen Brain Observatory Visual Coding dataset, we use representational similarity analysis to quantify the extent to which MouseNet recapitulates the neural representation in mouse visual cortex. Importantly, we provide evidence that optimizing for task performance does not improve similarity to the corresponding biological system beyond a certain point. We demonstrate that the distributions of some physiological quantities are closer to the observed distributions in the mouse brain after task training. We encourage the use of the MouseNet architecture by making the code freely available.

What is So Special About Contemporary CG Faces? Semiotics of MetaHumans

This paper analyses the features of the 2021 software for the creation of ultrarealistic digital characters "MetaHuman Creator" and reflects on the causes of such perceived effect of realism to understand if the faces produced with such software represent an actual novelty from an academic standpoint. Such realism is first of all defined as the result of semio-cognitive processes which trigger interpretative habits specifically related to faces. These habits are then related to the main properties of any realistic face: being face-looking, face-meaning and face-acting. These properties, in turn, are put in relation with our interactions with faces in terms of face detection, face recognition, face reading and face agency. Within this theoretical framework, we relate the characteristics of these artificial faces with such interpretative habits. To do so, we first of all make an examination of the technological features behind both the software and the digital faces it produces. This analysis highlights four main points of interest: the mathematical accuracy, the scanned database, the high level of details and the transformative capacities of these artificial faces. We then relate these characteristics with the cultural and cognitive aspects involved in recognizing and granting meaning to faces. This reveals how metahuman faces differs from previous artificial faces in terms of indexicality, intersubjectivity, informativity and irreducibility. But it also reveals some limits of such effect of reality in terms of intentionality and historical context. This examination consequently brings us to conclude that metahuman faces are qualitatively different from previous artificial faces and, in the light of their potentials and limits, to highlight four main lines of future research based on our findings.

Automatic Segmentation and Quantitative Assessment of Stroke Lesions on MR Images

Lesion studies are crucial in establishing brain-behavior relationships, and accurately segmenting the lesion represents the first step in achieving this. Manual lesion segmentation is the gold standard for chronic strokes. However, it is labor-intensive, subject to bias, and limits sample size. Therefore, our objective is to develop an automatic segmentation algorithm for chronic stroke lesions on T1-weighted MR images. Methods: To train our model, we utilized an open-source dataset: ATLAS v2.0 (Anatomical Tracings of Lesions After Stroke). We partitioned the dataset of 655 T1 images with manual segmentation labels into five subsets and performed a 5-fold cross-validation to avoid overfitting of the model. We used a deep neural network (DNN) architecture for model training. Results: To evaluate the model performance, we used three metrics that pertain to diverse aspects of volumetric segmentation, including shape, location, and size. The Dice similarity coefficient (DSC) compares the spatial overlap between manual and machine segmentation. The average DSC was 0.65 (0.61−0.67; 95% bootstrapped CI). Average symmetric surface distance (ASSD) measures contour distances between the two segmentations. ASSD between manual and automatic segmentation was 12 mm. Finally, we compared the total lesion volumes and the Pearson correlation coefficient (ρ) between the manual and automatically segmented lesion volumes, which was 0.97 (p-value < 0.001). Conclusions: We present the first automated segmentation model trained on a large multicentric dataset. This model will enable automated on-demand processing of MRI scans and quantitative chronic stroke lesion assessment.

Precision dynamical mapping using topological data analysis reveals a hub-like transition state at rest

In the absence of external stimuli, neural activity continuously evolves from one configuration to another. Whether these transitions or explorations follow some underlying arrangement or lack a predictable ordered plan remains to be determined. Here, using fMRI data from highly sampled individuals (~5 hours of resting-state data per individual), we aimed to reveal the rules that govern transitions in brain activity at rest. Our Topological Data Analysis based Mapper approach characterized a highly visited transition state of the brain that acts as a switch between different neural configurations to organize the spontaneous brain activity. Further, while the transition state was characterized by a uniform representation of canonical resting-state networks (RSNs), the periphery of the landscape was dominated by a subject-specific combination of RSNs. Altogether, we revealed rules or principles that organize spontaneous brain activity using a precision dynamics approach.

Real-world size of objects serves as an axis of object space

Our mind can represent various objects from physical world in an abstract and complex high-dimensional object space, with axes encoding critical features to quickly and accurately recognize objects. Among object features identified in previous neurophysiological and fMRI studies that may serve as the axes, objects’ real-world size is of particular interest because it provides not only visual information for broad conceptual distinctions between objects but also ecological information for objects’ affordance. Here we use deep convolutional neural networks (DCNNs), which enable direct manipulation of visual experience and units’ activation, to explore how objects’ real-world size is extracted to construct the axis of object space. Like the human brain, the DCNNs pre-trained for object recognition also encode objects’ size as an independent axis of the object space. Further, we find that the shape of objects, rather than retinal size, context, task demands or texture features, is critical to inferring objects’ size for both DCNNs and humans. In short, with DCNNs as a brain-like model, our study devises a paradigm supplemental to conventional approaches to explore the structure of object space, which provides computational support for empirical observations on human perceptual and neural representations of objects.

Combined fMRI and deep convolutional neural network analysis suggest that an axis of object space specifically encodes objects’ real-world size based on objects’ shape.

Unraveling Functional Diversity of Cortical Synaptic Architecture Through the Lens of Population Coding

The synaptic inputs to single cortical neurons exhibit substantial diversity in their sensory-driven activity. What this diversity reflects is unclear, and appears counter-productive in generating selective somatic responses to specific stimuli. One possibility is that this diversity reflects the propagation of information from one neural population to another. To test this possibility, we bridge population coding theory with measurements of synaptic inputs recorded in vivo with two-photon calcium imaging. We construct a probabilistic decoder to estimate the stimulus orientation from the responses of a realistic, hypothetical input population of neurons to compare with synaptic inputs onto individual neurons of ferret primary visual cortex (V1) recorded with two-photon calcium imaging in vivo. We find that optimal decoding requires diverse input weights and provides a straightforward mapping from the decoder weights to excitatory synapses. Analytically derived weights for biologically realistic input populations closely matched the functional heterogeneity of dendritic spines imaged in vivo with two-photon calcium imaging. Our results indicate that synaptic diversity is a necessary component of information transmission and reframes studies of connectivity through the lens of probabilistic population codes. These results suggest that the mapping from synaptic inputs to somatic selectivity may not be directly interpretable without considering input covariance and highlights the importance of population codes in pursuit of the cortical connectome.

Capacity for movement is an organisational principle in object representations

The ability to perceive moving objects is crucial for threat identification and survival. Recent neuroimaging evidence has shown that goal-directed movement is an important element of object processing in the brain. However, prior work has primarily used moving stimuli that are also animate, making it difficult to disentangle the effect of movement from aliveness or animacy in representational categorisation. In the current study, we investigated the relationship between how the brain processes movement and aliveness by including stimuli that are alive but still (e.g., plants), and stimuli that are not alive but move (e.g., waves). We examined electroencephalographic (EEG) data recorded while participants viewed static images of moving or non-moving objects that were either natural or artificial. Participants classified the images according to aliveness, or according to capacity for movement. Movement explained significant variance in the neural data over and above that of aliveness, showing that capacity for movement is an important dimension in the representation of visual objects in humans.

Reassessing hierarchical correspondences between brain and deep networks through direct interface

Functional correspondences between deep convolutional neural networks (DCNNs) and the mammalian visual system support a hierarchical account in which successive stages of processing contain ever higher-level information. However, these correspondences between brain and model activity involve shared, not task-relevant, variance. We propose a stricter account of correspondence: If a DCNN layer corresponds to a brain region, then replacing model activity with brain activity should successfully drive the DCNN's object recognition decision. Using this approach on three datasets, we found that all regions along the ventral visual stream best corresponded with later model layers, indicating that all stages of processing contained higher-level information about object category. Time course analyses suggest that long-range recurrent connections transmit object class information from late to early visual areas.

Recurrent Connections in the Primate Ventral Visual Stream Mediate a Trade-Off Between Task Performance and Network Size During Core Object Recognition

The computational role of the abundant feedback connections in the ventral visual stream is unclear, enabling humans and nonhuman primates to effortlessly recognize objects across a multitude of viewing conditions. Prior studies have augmented feedforward convolutional neural networks (CNNs) with recurrent connections to study their role in visual processing; however, often these recurrent networks are optimized directly on neural data or the comparative metrics used are undefined for standard feedforward networks that lack these connections. In this work, we develop task-optimized convolutional recurrent (ConvRNN) network models that more correctly mimic the timing and gross neuroanatomy of the ventral pathway. Properly chosen intermediate-depth ConvRNN circuit architectures, which incorporate mechanisms of feedforward bypassing and recurrent gating, can achieve high performance on a core recognition task, comparable to that of much deeper feedforward networks. We then develop methods that allow us to compare both CNNs and ConvRNNs to finely grained measurements of primate categorization behavior and neural response trajectories across thousands of stimuli. We find that high-performing ConvRNNs provide a better match to these data than feedforward networks of any depth, predicting the precise timings at which each stimulus is behaviorally decoded from neural activation patterns. Moreover, these ConvRNN circuits consistently produce quantitatively accurate predictions of neural dynamics from V4 and IT across the entire stimulus presentation. In fact, we find that the highest-performing ConvRNNs, which best match neural and behavioral data, also achieve a strong Pareto trade-off between task performance and overall network size. Taken together, our results suggest the functional purpose of recurrence in the ventral pathway is to fit a high-performing network in cortex, attaining computational power through temporal rather than spatial complexity.

Learning in deep neural networks and brains with similarity-weighted interleaved learning

Understanding how the brain learns throughout a lifetime remains a long-standing challenge. In artificial neural networks (ANNs), incorporating novel information too rapidly results in catastrophic interference, i.e., abrupt loss of previously acquired knowledge. Complementary Learning Systems Theory (CLST) suggests that new memories can be gradually integrated into the neocortex by interleaving new memories with existing knowledge. This approach, however, has been assumed to require interleaving all existing knowledge every time something new is learned, which is implausible because it is time-consuming and requires a large amount of data. We show that deep, nonlinear ANNs can learn new information by interleaving only a subset of old items that share substantial representational similarity with the new information. By using such similarity-weighted interleaved learning (SWIL), ANNs can learn new information rapidly with a similar accuracy level and minimal interference, while using a much smaller number of old items presented per epoch (fast and data-efficient). SWIL is shown to work with various standard classification datasets (Fashion-MNIST, CIFAR10, and CIFAR100), deep neural network architectures, and in sequential learning frameworks. We show that data efficiency and speedup in learning new items are increased roughly proportionally to the number of nonoverlapping classes stored in the network, which implies an enormous possible speedup in human brains, which encode a high number of separate categories. Finally, we propose a theoretical model of how SWIL might be implemented in the brain.

Trailblazers in Neuroscience: Using compositionality to understand how parts combine in whole objects

A fundamental question for any visual system is whether its image representation can be understood in terms of its components. Decomposing any image into components is challenging because there are many possible decompositions with no common dictionary, and enumerating them leads to a combinatorial explosion. Even in perception, many objects are readily seen as containing parts, but there are many exceptions. These exceptions include objects that are not perceived as containing parts, properties like symmetry that cannot be localized to any single part, and also special categories like words and faces whose perception is widely believed to be holistic. Here, I describe a novel approach we have used to address these issues and evaluate compositionality at the behavioral and neural levels. The key design principle is to create a large number of objects by combining a small number of pre-defined components in all possible ways. This allows for building component-based models that explain whole objects using a combination of these components. Importantly, any systematic error in model fits can be used to detect the presence of emergent or holistic properties. Using this approach, we have found that whole object representations are surprisingly predictable from their components, that some components are preferred to others in perception, and that emergent properties can be discovered or explained using compositional models. Thus, compositionality is a powerful approach for understanding how whole objects relate to their parts.

Statistical Learning in Vision

Vision and learning have long been considered to be two areas of research linked only distantly. However, recent developments in vision research have changed the conceptual definition of vision from a signal-evaluating process to a goal-oriented interpreting process, and this shift binds learning, together with the resulting internal representations, intimately to vision. In this review, we consider various types of learning (perceptual, statistical, and rule/abstract) associated with vision in the past decades and argue that they represent differently specialized versions of the fundamental learning process, which must be captured in its entirety when applied to complex visual processes. We show why the generalized version of statistical learning can provide the appropriate setup for such a unified treatment of learning in vision, what computational framework best accommodates this kind of statistical learning, and what plausible neural scheme could feasibly implement this framework. Finally, we list the challenges that the field of statistical learning faces in fulfilling the promise of being the right vehicle for advancing our understanding of vision in its entirety. Expected final online publication date for the Annual Review of Vision Science, Volume 8 is September 2022. Please see http://www.annualreviews.org/page/journal/pubdates for revised estimates.

Face identity coding in the deep neural network and primate brain

A central challenge in face perception research is to understand how neurons encode face identities. This challenge has not been met largely due to the lack of simultaneous access to the entire face processing neural network and the lack of a comprehensive multifaceted model capable of characterizing a large number of facial features. Here, we addressed this challenge by conducting in silico experiments using a pre-trained face recognition deep neural network (DNN) with a diverse array of stimuli. We identified a subset of DNN units selective to face identities, and these identity-selective units demonstrated generalized discriminability to novel faces. Visualization and manipulation of the network revealed the importance of identity-selective units in face recognition. Importantly, using our monkey and human single-neuron recordings, we directly compared the response of artificial units with real primate neurons to the same stimuli and found that artificial units shared a similar representation of facial features as primate neurons. We also observed a region-based feature coding mechanism in DNN units as in human neurons. Together, by directly linking between artificial and primate neural systems, our results shed light on how the primate brain performs face recognition tasks.

Identifying Objects and Remembering Images: Insights From Deep Neural Networks

People have a remarkable ability to identify the objects that they are looking at, as well as remember the images that they have seen. Researchers know that high-level visual cortex contributes in important ways to supporting both of these functions, but developing models that describe how processing in high-level visual cortex supports these behaviors has been challenging. Recent breakthroughs in this modeling effort have arrived by way of the illustration that deep artificial neural networks trained to categorize objects, developed for computer vision purposes, reflect brainlike patterns of activity. Here we summarize how deep artificial neural networks have been used to gain important insights into the contributions of high-level visual cortex to object identification, as well as one characteristic of visual memory behavior: image memorability, the systematic variation with which some images are remembered better than others.

Sensitivity of Sparse Codes to Image Distortions

Sparse coding has been proposed as a theory of visual cortex and as an unsupervised algorithm for learning representations. We show empirically with the MNIST data set that sparse codes can be very sensitive to image distortions, a behavior that may hinder invariant object recognition. A locally linear analysis suggests that the sensitivity is due to the existence of linear combinations of active dictionary elements with high cancellation. A nearest-neighbor classifier is shown to perform worse on sparse codes than original images. For a linear classifier with a sufficiently large number of labeled examples, sparse codes are shown to yield higher accuracy than original images, but no higher than a representation computed by a random feedforward net. Sensitivity to distortions seems to be a basic property of sparse codes, and one should be aware of this property when applying sparse codes to invariant object recognition.

Seeing the Forest for the Trees, and the Ground Below My Beak: Global and Local Processing in the Pigeon's Visual System

Non-human animals tend to solve behavioral tasks using local information. Pigeons are particularly biased toward using the local features of stimuli to guide behavior in small-scale environments. When behavioral tasks are performed in large-scale environments, pigeons are much better global processors of information. The local and global strategies are mediated by two different fovea in the pigeon retina that are associated with the tectofugal and thalamofugal pathways. We discuss the neural mechanisms of pigeons' bias for local information within the tectofugal pathway, which terminates at an intermediate stage of extracting shape complexity. We also review the evidence suggesting that the thalamofugal pathway participates in global processing in pigeons and is primarily engaged in constructing a spatial representation of the environment in conjunction with the hippocampus.

Twenty years of investigation with the case of prosopagnosia PS to understand human face identity recognition. Part II: Neural basis

Patient PS sustained her dramatic brain injury in 1992, the same year as the first report of a neuroimaging study of human face recognition. The present paper complements the review on the functional nature of PS's prosopagnosia (part I), illustrating how her case study directly, i.e., through neuroimaging investigations of her brain structure and activity, but also indirectly, through neural studies performed on other clinical cases and neurotypical individuals, inspired and constrained neural models of human face recognition. In the dominant right hemisphere for face recognition in humans, PS's main lesion concerns (inputs to) the inferior occipital gyrus (IOG), in a region where face-selective activity is typically found in normal individuals ('Occipital Face Area', OFA). Her case study initially supported the criticality of this region for face identity recognition (FIR) and provided the impetus for transcranial magnetic stimulation (TMS), intracerebral electrical stimulation, and cortical surgery studies that have generally supported this view. Despite PS's right IOG lesion, typical face-selectivity is found anteriorly in the middle portion of the fusiform gyrus, a hominoid structure (termed the right 'Fusiform Face Area', FFA) that is widely considered to be the most important region for human face recognition. This finding led to the original proposal of direct anatomico-functional connections from early visual cortices to the FFA, bypassing the IOG/OFA (lulu), a hypothesis supported by further neuroimaging studies of PS, other neurological cases and neuro-typical individuals with original visual stimulation paradigms, data recordings and analyses. The proposal of a lack of sensitivity to face identity in PS's right FFA due to defective reentrant inputs from the IOG/FFA has also been supported by other cases, functional connectivity and cortical surgery studies. Overall, neural studies of, and based on, the case of prosopagnosia PS strongly question the hierarchical organization of the human neural face recognition system, supporting a more flexible and dynamic view of this key social brain function.