A deep learning framework for neuroscience

Visual looming is a primitive for human emotion

The neural computations for looming detection are strikingly similar across species. In mammals, information about approaching threats is conveyed from the retina to the midbrain superior colliculus, where approach variables are computed to enable defensive behavior. Although neuroscientific theories posit that midbrain representations contribute to emotion through connectivity with distributed brain systems, it remains unknown whether a computational system for looming detection can predict both defensive behavior and phenomenal experience in humans. Here, we show that a shallow convolutional neural network based on the Drosophila visual system predicts defensive blinking to looming objects in infants and superior colliculus responses to optical expansion in adults. Further, the neural network's responses to naturalistic video clips predict self-reported emotion largely by way of subjective arousal. These findings illustrate how a simple neural network architecture optimized for a species-general task relevant for survival explains motor and experiential components of human emotion.

Contribution of low-level image statistics to EEG decoding of semantic content in multivariate and univariate models with feature optimization

Spatio-temporal patterns of evoked brain activity contain information that can be used to decode and categorize the semantic content of visual stimuli. However, this procedure can be biased by low-level image features independently of the semantic content present in the stimuli, prompting the need to understand the robustness of different models regarding these confounding factors. In this study, we trained machine learning models to distinguish between concepts included in the publicly available THINGS-EEG dataset using electroencephalography (EEG) data acquired during a rapid serial visual presentation paradigm. We investigated the contribution of low-level image features to decoding accuracy in a multivariate model, utilizing broadband data from all EEG channels. Additionally, we explored a univariate model obtained through data-driven feature selection applied to the spatial and frequency domains. While the univariate models exhibited better decoding accuracy, their predictions were less robust to the confounding effect of low-level image statistics. Notably, some of the models maintained their accuracy even after random replacement of the training dataset with semantically unrelated samples that presented similar low-level content. In conclusion, our findings suggest that model optimization impacts sensitivity to confounding factors, regardless of the resulting classification performance. Therefore, the choice of EEG features for semantic decoding should ideally be informed by criteria beyond classifier performance, such as the neurobiological mechanisms under study.

How well do models of visual cortex generalize to out of distribution samples?

Unit activity in particular deep neural networks (DNNs) are remarkably similar to the neuronal population responses to static images along the primate ventral visual cortex. Linear combinations of DNN unit activities are widely used to build predictive models of neuronal activity in the visual cortex. Nevertheless, prediction performance in these models is often investigated on stimulus sets consisting of everyday objects under naturalistic settings. Recent work has revealed a generalization gap in how predicting neuronal responses to synthetically generated out-of-distribution (OOD) stimuli. Here, we investigated how the recent progress in improving DNNs' object recognition generalization, as well as various DNN design choices such as architecture, learning algorithm, and datasets have impacted the generalization gap in neural predictivity. We came to a surprising conclusion that the performance on none of the common computer vision OOD object recognition benchmarks is predictive of OOD neural predictivity performance. Furthermore, we found that adversarially robust models often yield substantially higher generalization in neural predictivity, although the degree of robustness itself was not predictive of neural predictivity score. These results suggest that improving object recognition behavior on current benchmarks alone may not lead to more general models of neurons in the primate ventral visual cortex.

Temporal dynamics of short-term neural adaptation across human visual cortex

Neural responses in visual cortex adapt to prolonged and repeated stimuli. While adaptation occurs across the visual cortex, it is unclear how adaptation patterns and computational mechanisms differ across the visual hierarchy. Here we characterize two signatures of short-term neural adaptation in time-varying intracranial electroencephalography (iEEG) data collected while participants viewed naturalistic image categories varying in duration and repetition interval. Ventral- and lateral-occipitotemporal cortex exhibit slower and prolonged adaptation to single stimuli and slower recovery from adaptation to repeated stimuli compared to V1-V3. For category-selective electrodes, recovery from adaptation is slower for preferred than non-preferred stimuli. To model neural adaptation we augment our delayed divisive normalization (DN) model by scaling the input strength as a function of stimulus category, enabling the model to accurately predict neural responses across multiple image categories. The model fits suggest that differences in adaptation patterns arise from slower normalization dynamics in higher visual areas interacting with differences in input strength resulting from category selectivity. Our results reveal systematic differences in temporal adaptation of neural population responses between lower and higher visual brain areas and show that a single computational model of history-dependent normalization dynamics, fit with area-specific parameters, accounts for these differences.

Philosophy of cognitive science in the age of deep learning

Deep learning has enabled major advances across most areas of artificial intelligence research. This remarkable progress extends beyond mere engineering achievements and holds significant relevance for the philosophy of cognitive science. Deep neural networks have made significant strides in overcoming the limitations of older connectionist models that once occupied the center stage of philosophical debates about cognition. This development is directly relevant to long-standing theoretical debates in the philosophy of cognitive science. Furthermore, ongoing methodological challenges related to the comparative evaluation of deep neural networks stand to benefit greatly from interdisciplinary collaboration with philosophy and cognitive science. The time is ripe for philosophers to explore foundational issues related to deep learning and cognition; this perspective paper surveys key areas where their contributions can be especially fruitful. This article is categorized under: Philosophy > Artificial Intelligence Computer Science and Robotics > Machine Learning.

Artificial Intelligence and Colposcopy: Automatic Identification of Cervical Squamous Cell Carcinoma Precursors

Background/Objectives: Proficient colposcopy is crucial for the adequate management of cervical cancer precursor lesions; nonetheless its limitations may impact its cost-effectiveness. The development of artificial intelligence models is experiencing an exponential growth, particularly in image-based specialties. The aim of this study is to develop and validate a Convolutional Neural Network (CNN) for the automatic differentiation of high-grade (HSIL) from low-grade dysplasia (LSIL) in colposcopy. Methods: A unicentric retrospective study was conducted based on 70 colposcopy exams, comprising a total of 22,693 frames. Among these, 8729 were categorized as HSIL based on histopathology. The total dataset was divided into a training (90%, n = 20,423) and a testing set (10%, n = 2270), the latter being used to evaluate the model's performance. The main outcome measures included sensitivity, specificity, accuracy, positive predictive value (PPV), negative predictive value (NPV), and the area under the receiving operating curve (AUC-ROC). Results: The sensitivity was 99.7% and the specificity was 98.6%. The PPV and NPV were 97.8% and 99.8%, respectively. The overall accuracy was 99.0%. The AUC-ROC was 0.98. The CNN processed 112 frames per second. Conclusions: We developed a CNN capable of differentiating cervical cancer precursors in colposcopy frames. The high levels of accuracy for the differentiation of HSIL from LSIL may improve the diagnostic yield of this exam.

Investigating the discrimination ability of 3D convolutional neural networks applied to altered brain MRI parametric maps

Convolutional neural networks (CNNs) are gradually being recognized in the neuroimaging community as a powerful tool for image analysis. Despite their outstanding performances, some aspects of CNN functioning are still not fully understood by human operators. We postulated that the interpretability of CNNs applied to neuroimaging data could be improved by investigating their behavior when they are fed data with known characteristics. We analyzed the ability of 3D CNNs to discriminate between original and altered whole-brain parametric maps derived from diffusion-weighted magnetic resonance imaging. The alteration consisted in linearly changing the voxel intensity of either one (monoregion) or two (biregion) anatomical regions in each brain volume, but without mimicking any neuropathology. Performing ten-fold cross-validation and using a hold-out set for testing, we assessed the CNNs' discrimination ability according to the intensity of the altered regions, comparing the latter's size and relative position. Monoregion CNNs showed that the larger the modified region, the smaller the intensity increase needed to achieve good performances. Biregion CNNs systematically outperformed monoregion CNNs, but could only detect one of the two target regions when tested on the corresponding monoregion images. Exploiting prior information on training data allowed for a better understanding of CNN behavior, especially when altered regions were combined. This can inform about the complexity of CNN pattern retrieval and elucidate misclassified examples, particularly relevant for pathological data. The proposed analytical approach may serve to gain insights into CNN behavior and guide the design of enhanced detection systems exploiting our prior knowledge.

Transition to chaos separates learning regimes and relates to measure of consciousness in recurrent neural networks

Recurrent neural networks exhibit chaotic dynamics when the variance in their connection strengths exceed a critical value. Recent work indicates connection variance also modulates learning strategies; networks learn "rich" representations when initialized with low coupling and "lazier"solutions with larger variance. Using Watts-Strogatz networks of varying sparsity, structure, and hidden weight variance, we find that the critical coupling strength dividing chaotic from ordered dynamics also differentiates rich and lazy learning strategies. Training moves both stable and chaotic networks closer to the edge of chaos, with networks learning richer representations before the transition to chaos. In contrast, biologically realistic connectivity structures foster stability over a wide range of variances. The transition to chaos is also reflected in a measure that clinically discriminates levels of consciousness, the perturbational complexity index (PCIst). Networks with high values of PCIst exhibit stable dynamics and rich learning, suggesting a consciousness prior may promote rich learning. The results suggest a clear relationship between critical dynamics, learning regimes and complexity-based measures of consciousness.

A unifying framework for functional organization in early and higher ventral visual cortex

A key feature of cortical systems is functional organization: the arrangement of functionally distinct neurons in characteristic spatial patterns. However, the principles underlying the emergence of functional organization in the cortex are poorly understood. Here, we develop the topographic deep artificial neural network (TDANN), the first model to predict several aspects of the functional organization of multiple cortical areas in the primate visual system. We analyze the factors driving the TDANN's success and find that it balances two objectives: learning a task-general sensory representation and maximizing the spatial smoothness of responses according to a metric that scales with cortical surface area. In turn, the representations learned by the TDANN are more brain-like than in spatially unconstrained models. Finally, we provide evidence that the TDANN's functional organization balances performance with between-area connection length. Our results offer a unified principle for understanding the functional organization of the primate ventral visual system.

A new era in cognitive neuroscience: the tidal wave of artificial intelligence (AI)

Translating artificial intelligence techniques into the realm of cognitive neuroscience holds promise for significant breakthroughs in our ability to probe the intrinsic mechanisms of the brain. The recent unprecedented development of robust AI models is changing how and what we understand about the brain. In this Editorial, we invite contributions for a BMC Neuroscience Collection on "AI and Cognitive Neuroscience".

Representational drift as a result of implicit regularization

Recent studies show that, even in constant environments, the tuning of single neurons changes over time in a variety of brain regions. This representational drift has been suggested to be a consequence of continuous learning under noise, but its properties are still not fully understood. To investigate the underlying mechanism, we trained an artificial network on a simplified navigational task. The network quickly reached a state of high performance, and many units exhibited spatial tuning. We then continued training the network and noticed that the activity became sparser with time. Initial learning was orders of magnitude faster than ensuing sparsification. This sparsification is consistent with recent results in machine learning, in which networks slowly move within their solution space until they reach a flat area of the loss function. We analyzed four datasets from different labs, all demonstrating that CA1 neurons become sparser and more spatially informative with exposure to the same environment. We conclude that learning is divided into three overlapping phases: (i) Fast familiarity with the environment; (ii) slow implicit regularization; and (iii) a steady state of null drift. The variability in drift dynamics opens the possibility of inferring learning algorithms from observations of drift statistics.

Computational limits to the legibility of the imaged human brain

Our knowledge of the organisation of the human brain at the population-level is yet to translate into power to predict functional differences at the individual-level, limiting clinical applications and casting doubt on the generalisability of inferred mechanisms. It remains unknown whether the difficulty arises from the absence of individuating biological patterns within the brain, or from limited power to access them with the models and compute at our disposal. Here we comprehensively investigate the resolvability of such patterns with data and compute at unprecedented scale. Across 23 810 unique participants from UK Biobank, we systematically evaluate the predictability of 25 individual biological characteristics, from all available combinations of structural and functional neuroimaging data. Over 4526 GPU*hours of computation, we train, optimize, and evaluate out-of-sample 700 individual predictive models, including fully-connected feed-forward neural networks of demographic, psychological, serological, chronic disease, and functional connectivity characteristics, and both uni- and multi-modal 3D convolutional neural network models of macro- and micro-structural brain imaging. We find a marked discrepancy between the high predictability of sex (balanced accuracy 99.7%), age (mean absolute error 2.048 years, R2 0.859), and weight (mean absolute error 2.609Kg, R2 0.625), for which we set new state-of-the-art performance, and the surprisingly low predictability of other characteristics. Neither structural nor functional imaging predicted an individual's psychology better than the coincidence of common chronic disease (p < 0.05). Serology predicted chronic disease (p < 0.05) and was best predicted by it (p < 0.001), followed by structural neuroimaging (p < 0.05). Our findings suggest either more informative imaging or more powerful models will be needed to decipher individual level characteristics from the human brain. We make our models and code openly available.

Mapping model units to visual neurons reveals population code for social behaviour

The rich variety of behaviours observed in animals arises through the interplay between sensory processing and motor control. To understand these sensorimotor transformations, it is useful to build models that predict not only neural responses to sensory input1-5 but also how each neuron causally contributes to behaviour6,7. Here we demonstrate a novel modelling approach to identify a one-to-one mapping between internal units in a deep neural network and real neurons by predicting the behavioural changes that arise from systematic perturbations of more than a dozen neuronal cell types. A key ingredient that we introduce is 'knockout training', which involves perturbing the network during training to match the perturbations of the real neurons during behavioural experiments. We apply this approach to model the sensorimotor transformations of Drosophila melanogaster males during a complex, visually guided social behaviour8-11. The visual projection neurons at the interface between the optic lobe and central brain form a set of discrete channels12, and prior work indicates that each channel encodes a specific visual feature to drive a particular behaviour13,14. Our model reaches a different conclusion: combinations of visual projection neurons, including those involved in non-social behaviours, drive male interactions with the female, forming a rich population code for behaviour. Overall, our framework consolidates behavioural effects elicited from various neural perturbations into a single, unified model, providing a map from stimulus to neuronal cell type to behaviour, and enabling future incorporation of wiring diagrams of the brain15 into the model.

Spiking neural networks for nonlinear regression

Spiking neural networks (SNN), also often referred to as the third generation of neural networks, carry the potential for a massive reduction in memory and energy consumption over traditional, second-generation neural networks. Inspired by the undisputed efficiency of the human brain, they introduce temporal and neuronal sparsity, which can be exploited by next-generation neuromorphic hardware. Energy efficiency plays a crucial role in many engineering applications, for instance, in structural health monitoring. Machine learning in engineering contexts, especially in data-driven mechanics, focuses on regression. While regression with SNN has already been discussed in a variety of publications, in this contribution, we provide a novel formulation for its accuracy and energy efficiency. In particular, a network topology for decoding binary spike trains to real numbers is introduced, using the membrane potential of spiking neurons. Several different spiking neural architectures, ranging from simple spiking feed-forward to complex spiking long short-term memory neural networks, are derived. Since the proposed architectures do not contain any dense layers, they exploit the full potential of SNN in terms of energy efficiency. At the same time, the accuracy of the proposed SNN architectures is demonstrated by numerical examples, namely different material models. Linear and nonlinear, as well as history-dependent material models, are examined. While this contribution focuses on mechanical examples, the interested reader may regress any custom function by adapting the published source code.

Artificial neural network identified a 20-gene panel in predicting immunotherapy response and survival benefits after anti-PD1/PD-L1 treatment in glioblastoma patients

BACKGROUND

Immune checkpoint inhibitors (ICIs) are a promising immunotherapy approach, but glioblastoma clinical trials have not yielded satisfactory results.

OBJECTIVE

To screen glioblastoma patients who may benefit from immunotherapy.

METHODS

Eighty-one patients receiving anti-PD1/PD-L1 treatment from a large-scale clinical trial and 364 patients without immunotherapy from The Cancer Genome Atlas (TCGA) were included. Patients in the ICI-treated cohort were divided into responders and nonresponders according to overall survival (OS), and the most critical responder-relevant features were screened using random forest (RF). We constructed an artificial neural network (ANN) model and verified its predictive value with immunotherapy response and OS.

RESULTS

We defined two groups of ICI-treated glioblastoma patients with large differences in survival benefits as nonresponders (OS ≤6 months, n = 18) and responders (OS ≥17 months, n = 8). No differentially mutated genes were observed between responders and nonresponders. We performed RF analysis to select the most critical responder-relevant features and developed an ANN with 20 input variables, five hidden neurons and one output neuron. Receiver operating characteristic analysis and the DeLong test demonstrated that the ANN had the best performance in predicting responders, with an AUC of 0.97. Survival analysis indicated that ANN-predicted responders had significantly better OS rates than nonresponders.

CONCLUSION

The 20-gene panel developed by the ANN could be a promising biomarker for predicting immunotherapy response and prognostic benefits in ICI-treated GBM patients and may guide oncologists to accurately select potential responders for the preferential use of ICIs.

Neuromorphic one-shot learning utilizing a phase-transition material

Design of hardware based on biological principles of neuronal computation and plasticity in the brain is a leading approach to realizing energy- and sample-efficient AI and learning machines. An important factor in selection of the hardware building blocks is the identification of candidate materials with physical properties suitable to emulate the large dynamic ranges and varied timescales of neuronal signaling. Previous work has shown that the all-or-none spiking behavior of neurons can be mimicked by threshold switches utilizing material phase transitions. Here, we demonstrate that devices based on a prototypical metal-insulator-transition material, vanadium dioxide (VO2), can be dynamically controlled to access a continuum of intermediate resistance states. Furthermore, the timescale of their intrinsic relaxation can be configured to match a range of biologically relevant timescales from milliseconds to seconds. We exploit these device properties to emulate three aspects of neuronal analog computation: fast (~1 ms) spiking in a neuronal soma compartment, slow (~100 ms) spiking in a dendritic compartment, and ultraslow (~1 s) biochemical signaling involved in temporal credit assignment for a recently discovered biological mechanism of one-shot learning. Simulations show that an artificial neural network using properties of VO2 devices to control an agent navigating a spatial environment can learn an efficient path to a reward in up to fourfold fewer trials than standard methods. The phase relaxations described in our study may be engineered in a variety of materials and can be controlled by thermal, electrical, or optical stimuli, suggesting further opportunities to emulate biological learning in neuromorphic hardware.

A mathematical theory of relational generalization in transitive inference

Humans and animals routinely infer relations between different items or events and generalize these relations to novel combinations of items. This allows them to respond appropriately to radically novel circumstances and is fundamental to advanced cognition. However, how learning systems (including the brain) can implement the necessary inductive biases has been unclear. Here we investigated transitive inference (TI), a classic relational task paradigm in which subjects must learn a relation (A > B and B > C) and generalize it to new combinations of items (A > C). Through mathematical analysis, we found that a broad range of biologically relevant learning models (e.g. gradient flow or ridge regression) perform TI successfully and recapitulate signature behavioral patterns long observed in living subjects. First, we found that models with item-wise additive representations automatically encode transitive relations. Second, for more general representations, a single scalar "conjunctivity factor" determines model behavior on TI and, further, the principle of norm minimization (a standard statistical inductive bias) enables models with fixed, partly conjunctive representations to generalize transitively. Finally, neural networks in the "rich regime," which enables representation learning and has been found to improve generalization, unexpectedly show poor generalization and anomalous behavior. We find that such networks implement a form of norm minimization (over hidden weights) that yields a local encoding mechanism lacking transitivity. Our findings show how minimal statistical learning principles give rise to a classical relational inductive bias (transitivity), explain empirically observed behaviors, and establish a formal approach to understanding the neural basis of relational abstraction.

Uncovering the determinants of brain functioning, behavior and their interplay in the light of context

Notwithstanding the huge progress in molecular and cellular neuroscience, our ability to understand the brain and develop effective treatments promoting mental health is still limited. This can be partially ascribed to the reductionist, deterministic and mechanistic approaches in neuroscience that struggle with the complexity of the central nervous system. Here, I introduce the Context theory of constrained systems proposing a novel role of contextual factors and genetic, molecular and neural substrates in determining brain functioning and behavior. This theory entails key conceptual implications. First, context is the main driver of behavior and mental states. Second, substrates, from genes to brain areas, have no direct causal link to complex behavioral responses as they can be combined in multiple ways to produce the same response and different responses can impinge on the same substrates. Third, context and biological substrates play distinct roles in determining behavior: context drives behavior, substrates constrain the behavioral repertoire that can be implemented. Fourth, since behavior is the interface between the central nervous system and the environment, it is a privileged level of control and orchestration of brain functioning. Such implications are illustrated through the Kitchen metaphor of the brain. This theoretical framework calls for the revision of key concepts in neuroscience and psychiatry, including causality, specificity and individuality. Moreover, at the clinical level, it proposes treatments inducing behavioral changes through contextual interventions as having the highest impact to reorganize the complexity of the human mind and to achieve a long-lasting improvement in mental health.

Representational formats of human memory traces

Neural representations are internal brain states that constitute the brain's model of the external world or some of its features. In the presence of sensory input, a representation may reflect various properties of this input. When perceptual information is no longer available, the brain can still activate representations of previously experienced episodes due to the formation of memory traces. In this review, we aim at characterizing the nature of neural memory representations and how they can be assessed with cognitive neuroscience methods, mainly focusing on neuroimaging. We discuss how multivariate analysis techniques such as representational similarity analysis (RSA) and deep neural networks (DNNs) can be leveraged to gain insights into the structure of neural representations and their different representational formats. We provide several examples of recent studies which demonstrate that we are able to not only measure memory representations using RSA but are also able to investigate their multiple formats using DNNs. We demonstrate that in addition to slow generalization during consolidation, memory representations are subject to semantization already during short-term memory, by revealing a shift from visual to semantic format. In addition to perceptual and conceptual formats, we describe the impact of affective evaluations as an additional dimension of episodic memories. Overall, these studies illustrate how the analysis of neural representations may help us gain a deeper understanding of the nature of human memory.

Emergent neural dynamics and geometry for generalization in a transitive inference task

Relational cognition-the ability to infer relationships that generalize to novel combinations of objects-is fundamental to human and animal intelligence. Despite this importance, it remains unclear how relational cognition is implemented in the brain due in part to a lack of hypotheses and predictions at the levels of collective neural activity and behavior. Here we discovered, analyzed, and experimentally tested neural networks (NNs) that perform transitive inference (TI), a classic relational task (if A > B and B > C, then A > C). We found NNs that (i) generalized perfectly, despite lacking overt transitive structure prior to training, (ii) generalized when the task required working memory (WM), a capacity thought to be essential to inference in the brain, (iii) emergently expressed behaviors long observed in living subjects, in addition to a novel order-dependent behavior, and (iv) expressed different task solutions yielding alternative behavioral and neural predictions. Further, in a large-scale experiment, we found that human subjects performing WM-based TI showed behavior inconsistent with a class of NNs that characteristically expressed an intuitive task solution. These findings provide neural insights into a classical relational ability, with wider implications for how the brain realizes relational cognition.

Alignment of brain embeddings and artificial contextual embeddings in natural language points to common geometric patterns

Contextual embeddings, derived from deep language models (DLMs), provide a continuous vectorial representation of language. This embedding space differs fundamentally from the symbolic representations posited by traditional psycholinguistics. We hypothesize that language areas in the human brain, similar to DLMs, rely on a continuous embedding space to represent language. To test this hypothesis, we densely record the neural activity patterns in the inferior frontal gyrus (IFG) of three participants using dense intracranial arrays while they listened to a 30-minute podcast. From these fine-grained spatiotemporal neural recordings, we derive a continuous vectorial representation for each word (i.e., a brain embedding) in each patient. Using stringent zero-shot mapping we demonstrate that brain embeddings in the IFG and the DLM contextual embedding space have common geometric patterns. The common geometric patterns allow us to predict the brain embedding in IFG of a given left-out word based solely on its geometrical relationship to other non-overlapping words in the podcast. Furthermore, we show that contextual embeddings capture the geometry of IFG embeddings better than static word embeddings. The continuous brain embedding space exposes a vector-based neural code for natural language processing in the human brain.

Task-driven neural network models predict neural dynamics of proprioception

Proprioception tells the brain the state of the body based on distributed sensory neurons. Yet, the principles that govern proprioceptive processing are poorly understood. Here, we employ a task-driven modeling approach to investigate the neural code of proprioceptive neurons in cuneate nucleus (CN) and somatosensory cortex area 2 (S1). We simulated muscle spindle signals through musculoskeletal modeling and generated a large-scale movement repertoire to train neural networks based on 16 hypotheses, each representing different computational goals. We found that the emerging, task-optimized internal representations generalize from synthetic data to predict neural dynamics in CN and S1 of primates. Computational tasks that aim to predict the limb position and velocity were the best at predicting the neural activity in both areas. Since task optimization develops representations that better predict neural activity during active than passive movements, we postulate that neural activity in the CN and S1 is top-down modulated during goal-directed movements.

Deep CANALs: a deep learning approach to refining the canalization theory of psychopathology

Psychedelic therapy has seen a resurgence of interest in the last decade, with promising clinical outcomes for the treatment of a variety of psychopathologies. In response to this success, several theoretical models have been proposed to account for the positive therapeutic effects of psychedelics. One of the more prominent models is "RElaxed Beliefs Under pSychedelics," which proposes that psychedelics act therapeutically by relaxing the strength of maladaptive high-level beliefs encoded in the brain. The more recent "CANAL" model of psychopathology builds on the explanatory framework of RElaxed Beliefs Under pSychedelics by proposing that canalization (the development of overly rigid belief landscapes) may be a primary factor in psychopathology. Here, we make use of learning theory in deep neural networks to develop a series of refinements to the original CANAL model. Our primary theoretical contribution is to disambiguate two separate optimization landscapes underlying belief representation in the brain and describe the unique pathologies which can arise from the canalization of each. Along each dimension, we identify pathologies of either too much or too little canalization, implying that the construct of canalization does not have a simple linear correlation with the presentation of psychopathology. In this expanded paradigm, we demonstrate the ability to make novel predictions regarding what aspects of psychopathology may be amenable to psychedelic therapy, as well as what forms of psychedelic therapy may ultimately be most beneficial for a given individual.

Preliminary Evidence for Global Properties in Human Listeners During Natural Auditory Scene Perception

Theories of auditory and visual scene analysis suggest the perception of scenes relies on the identification and segregation of objects within it, resembling a detail-oriented processing style. However, a more global process may occur while analyzing scenes, which has been evidenced in the visual domain. It is our understanding that a similar line of research has not been explored in the auditory domain; therefore, we evaluated the contributions of high-level global and low-level acoustic information to auditory scene perception. An additional aim was to increase the field's ecological validity by using and making available a new collection of high-quality auditory scenes. Participants rated scenes on 8 global properties (e.g., open vs. enclosed) and an acoustic analysis evaluated which low-level features predicted the ratings. We submitted the acoustic measures and average ratings of the global properties to separate exploratory factor analyses (EFAs). The EFA of the acoustic measures revealed a seven-factor structure explaining 57% of the variance in the data, while the EFA of the global property measures revealed a two-factor structure explaining 64% of the variance in the data. Regression analyses revealed each global property was predicted by at least one acoustic variable (R2 = 0.33-0.87). These findings were extended using deep neural network models where we examined correlations between human ratings of global properties and deep embeddings of two computational models: an object-based model and a scene-based model. The results support that participants' ratings are more strongly explained by a global analysis of the scene setting, though the relationship between scene perception and auditory perception is multifaceted, with differing correlation patterns evident between the two models. Taken together, our results provide evidence for the ability to perceive auditory scenes from a global perspective. Some of the acoustic measures predicted ratings of global scene perception, suggesting representations of auditory objects may be transformed through many stages of processing in the ventral auditory stream, similar to what has been proposed in the ventral visual stream. These findings and the open availability of our scene collection will make future studies on perception, attention, and memory for natural auditory scenes possible.

The mediodorsal thalamus in executive control

Executive control, the ability to organize thoughts and action plans in real time, is a defining feature of higher cognition. Classical theories have emphasized cortical contributions to this process, but recent studies have reinvigorated interest in the role of the thalamus. Although it is well established that local thalamic damage diminishes cognitive capacity, such observations have been difficult to inform functional models. Recent progress in experimental techniques is beginning to enrich our understanding of the anatomical, physiological, and computational substrates underlying thalamic engagement in executive control. In this review, we discuss this progress and particularly focus on the mediodorsal thalamus, which regulates the activity within and across frontal cortical areas. We end with a synthesis that highlights frontal thalamocortical interactions in cognitive computations and discusses its functional implications in normal and pathological conditions.

Visual perception: On the trail of high-level shape aftereffects

The representation of visual shape is a critical component of our perception of the objects around us. A new study exploited shape aftereffects to reveal the high-dimensional space of geometric features our brains use to represent shape.

Towards a universal mechanism for successful deep learning

Recently, the underlying mechanism for successful deep learning (DL) was presented based on a quantitative method that measures the quality of a single filter in each layer of a DL model, particularly VGG-16 trained on CIFAR-10. This method exemplifies that each filter identifies small clusters of possible output labels, with additional noise selected as labels outside the clusters. This feature is progressively sharpened with each layer, resulting in an enhanced signal-to-noise ratio (SNR), which leads to an increase in the accuracy of the DL network. In this study, this mechanism is verified for VGG-16 and EfficientNet-B0 trained on the CIFAR-100 and ImageNet datasets, and the main results are as follows. First, the accuracy and SNR progressively increase with the layers. Second, for a given deep architecture, the maximal error rate increases approximately linearly with the number of output labels. Third, similar trends were obtained for dataset labels in the range [3, 1000], thus supporting the universality of this mechanism. Understanding the performance of a single filter and its dominating features paves the way to highly dilute the deep architecture without affecting its overall accuracy, and this can be achieved by applying the filter's cluster connections (AFCC).

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

Recent neuroimaging studies have shown that the visual cortex plays an important role in representing the affective significance of visual input. The origin of these affect-specific visual representations is debated: they are intrinsic to the visual system versus they arise through reentry from frontal emotion processing structures such as the amygdala. We examined this problem by combining convolutional neural network (CNN) models of the human ventral visual cortex pre-trained on ImageNet with two datasets of affective images. Our results show that (1) in all layers of the CNN models, there were artificial neurons that responded consistently and selectively to neutral, pleasant, or unpleasant images and (2) lesioning these neurons by setting their output to 0 or enhancing these neurons by increasing their gain led to decreased or increased emotion recognition performance respectively. These results support the idea that the visual system may have the intrinsic ability to represent the affective significance of visual input and suggest that CNNs offer a fruitful platform for testing neuroscientific theories.

The Emerging Science of Interacting Minds

For over a century, psychology has focused on uncovering mental processes of a single individual. However, humans rarely navigate the world in isolation. The most important determinants of successful development, mental health, and our individual traits and preferences arise from interacting with other individuals. Social interaction underpins who we are, how we think, and how we behave. Here we discuss the key methodological challenges that have limited progress in establishing a robust science of how minds interact and the new tools that are beginning to overcome these challenges. A deep understanding of the human mind requires studying the context within which it originates and exists: social interaction.

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

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

Bayesian inference is facilitated by modular neural networks with different time scales

Various animals, including humans, have been suggested to perform Bayesian inferences to handle noisy, time-varying external information. In performing Bayesian inference by the brain, the prior distribution must be acquired and represented by sampling noisy external inputs. However, the mechanism by which neural activities represent such distributions has not yet been elucidated. Our findings reveal that networks with modular structures, composed of fast and slow modules, are adept at representing this prior distribution, enabling more accurate Bayesian inferences. Specifically, the modular network that consists of a main module connected with input and output layers and a sub-module with slower neural activity connected only with the main module outperformed networks with uniform time scales. Prior information was represented specifically by the slow sub-module, which could integrate observed signals over an appropriate period and represent input means and variances. Accordingly, the neural network could effectively predict the time-varying inputs. Furthermore, by training the time scales of neurons starting from networks with uniform time scales and without modular structure, the above slow-fast modular network structure and the division of roles in which prior knowledge is selectively represented in the slow sub-modules spontaneously emerged. These results explain how the prior distribution for Bayesian inference is represented in the brain, provide insight into the relevance of modular structure with time scale hierarchy to information processing, and elucidate the significance of brain areas with slower time scales.

Temporal dynamics of short-term neural adaptation across human visual cortex

Neural responses in visual cortex adapt to prolonged and repeated stimuli. While adaptation occurs across the visual cortex, it is unclear how adaptation patterns and computational mechanisms differ across the visual hierarchy. Here we characterize two signatures of short-term neural adaptation in time-varying intracranial electroencephalography (iEEG) data collected while participants viewed naturalistic image categories varying in duration and repetition interval. Ventral- and lateral-occipitotemporal cortex exhibit slower and prolonged adaptation to single stimuli and slower recovery from adaptation to repeated stimuli compared to V1-V3. For category-selective electrodes, recovery from adaptation is slower for preferred than non-preferred stimuli. To model neural adaptation we augment our delayed divisive normalization (DN) model by scaling the input strength as a function of stimulus category, enabling the model to accurately predict neural responses across multiple image categories. The model fits suggest that differences in adaptation patterns arise from slower normalization dynamics in higher visual areas interacting with differences in input strength resulting from category selectivity. Our results reveal systematic differences in temporal adaptation of neural population responses across the human visual hierarchy and show that a single computational model of history-dependent normalization dynamics, fit with area-specific parameters, accounts for these differences.

Deep social neuroscience: the promise and peril of using artificial neural networks to study the social brain

This review offers an accessible primer to social neuroscientists interested in neural networks. It begins by providing an overview of key concepts in deep learning. It then discusses three ways neural networks can be useful to social neuroscientists: (i) building statistical models to predict behavior from brain activity; (ii) quantifying naturalistic stimuli and social interactions; and (iii) generating cognitive models of social brain function. These applications have the potential to enhance the clinical value of neuroimaging and improve the generalizability of social neuroscience research. We also discuss the significant practical challenges, theoretical limitations and ethical issues faced by deep learning. If the field can successfully navigate these hazards, we believe that artificial neural networks may prove indispensable for the next stage of the field's development: deep social neuroscience.

How connectivity structure shapes rich and lazy learning in neural circuits

In theoretical neuroscience, recent work leverages deep learning tools to explore how some network attributes critically influence its learning dynamics. Notably, initial weight distributions with small (resp. large) variance may yield a rich (resp. lazy) regime, where significant (resp. minor) changes to network states and representation are observed over the course of learning. However, in biology, neural circuit connectivity could exhibit a low-rank structure and therefore differs markedly from the random initializations generally used for these studies. As such, here we investigate how the structure of the initial weights -- in particular their effective rank -- influences the network learning regime. Through both empirical and theoretical analyses, we discover that high-rank initializations typically yield smaller network changes indicative of lazier learning, a finding we also confirm with experimentally-driven initial connectivity in recurrent neural networks. Conversely, low-rank initialization biases learning towards richer learning. Importantly, however, as an exception to this rule, we find lazier learning can still occur with a low-rank initialization that aligns with task and data statistics. Our research highlights the pivotal role of initial weight structures in shaping learning regimes, with implications for metabolic costs of plasticity and risks of catastrophic forgetting.

Representational drift as a result of implicit regularization

Recent studies show that, even in constant environments, the tuning of single neurons changes over time in a variety of brain regions. This representational drift has been suggested to be a consequence of continuous learning under noise, but its properties are still not fully understood. To investigate the underlying mechanism, we trained an artificial network on a simplified navigational task. The network quickly reached a state of high performance, and many units exhibited spatial tuning. We then continued training the network and noticed that the activity became sparser with time. Initial learning was orders of magnitude faster than ensuing sparsification. This sparsification is consistent with recent results in machine learning, in which networks slowly move within their solution space until they reach a flat area of the loss function. We analyzed four datasets from different labs, all demonstrating that CA1 neurons become sparser and more spatially informative with exposure to the same environment. We conclude that learning is divided into three overlapping phases: (i) Fast familiarity with the environment; (ii) slow implicit regularization; (iii) a steady state of null drift. The variability in drift dynamics opens the possibility of inferring learning algorithms from observations of drift statistics.

Human-like dissociations between confidence and accuracy in convolutional neural networks

Prior research has shown that manipulating stimulus energy by changing both stimulus contrast and variability results in confidence-accuracy dissociations in humans. Specifically, even when performance is matched, higher stimulus energy leads to higher confidence. The most common explanation for this effect is the positive evidence heuristic where confidence neglects evidence that disconfirms the choice. However, an alternative explanation is the signal-and-variance-increase hypothesis, according to which these dissociations arise from low-level changes in the separation and variance of perceptual representations. Because artificial neural networks lack built-in confidence heuristics, they can serve as a test for the necessity of confidence heuristics in explaining confidence-accuracy dissociations. Therefore, we tested whether confidence-accuracy dissociations induced by stimulus energy manipulations emerge naturally in convolutional neural networks (CNNs). We found that, across three different energy manipulations, CNNs produced confidence-accuracy dissociations similar to those found in humans. This effect was present for a range of CNN architectures from shallow 4-layer networks to very deep ones, such as VGG-19 and ResNet -50 pretrained on ImageNet. Further, we traced back the reason for the confidence-accuracy dissociations in all CNNs to the same signal-and-variance increase that has been proposed for humans: higher stimulus energy increased the separation and variance of the CNNs' internal representations leading to higher confidence even for matched accuracy. These findings cast doubt on the necessity of the positive evidence heuristic to explain human confidence and establish CNNs as promising models for adjudicating between low-level, stimulus-driven and high-level, cognitive explanations of human behavior.

A prefrontal-thalamic circuit encodes social information for social recognition

Social recognition encompasses encoding social information and distinguishing unfamiliar from familiar individuals to form social relationships. Although the medial prefrontal cortex (mPFC) is known to play a role in social behavior, how identity information is processed and by which route it is communicated in the brain remains unclear. Here we report that a ventral midline thalamic area, nucleus reuniens (Re) that has reciprocal connections with the mPFC, is critical for social recognition in male mice. In vivo single-unit recordings and decoding analysis reveal that neural populations in both mPFC and Re represent different social stimuli, however, mPFC coding capacity is stronger. We demonstrate that chemogenetic inhibitions of Re impair the mPFC-Re neural synchronization and the mPFC social coding. Projection pathway-specific inhibitions by optogenetics reveal that the reciprocal connectivity between the mPFC and the Re is necessary for social recognition. These results reveal an mPFC-thalamic circuit for social information processing.

Unsupervised learning of mid-level visual representations

Recently, a confluence between trends in neuroscience and machine learning has brought a renewed focus on unsupervised learning, where sensory processing systems learn to exploit the statistical structure of their inputs in the absence of explicit training targets or rewards. Sophisticated experimental approaches have enabled the investigation of the influence of sensory experience on neural self-organization and its synaptic bases. Meanwhile, novel algorithms for unsupervised and self-supervised learning have become increasingly popular both as inspiration for theories of the brain, particularly for the function of intermediate visual cortical areas, and as building blocks of real-world learning machines. Here we review some of these recent developments, placing them in historical context and highlighting some research lines that promise exciting breakthroughs in the near future.

An Explainable and Personalized Cognitive Reasoning Model Based on Knowledge Graph: Toward Decision Making for General Practice

General practice plays a prominent role in primary health care (PHC). However, evidence has shown that the quality of PHC is still unsatisfactory, and the accuracy of clinical diagnosis and treatment must be improved in China. Decision making tools based on artificial intelligence can help general practitioners diagnose diseases, but most existing research is not sufficiently scalable and explainable. An explainable and personalized cognitive reasoning model based on knowledge graph (CRKG) proposed in this article can provide personalized diagnosis, perform decision making in general practice, and simulate the mode of thinking of human beings utilizing patients' electronic health records (EHRs) and knowledge graph. Taking abdominal diseases as the application point, an abdominal disease knowledge graph is first constructed in a semiautomated manner. Then, the CRKG designed referring to dual process theory in cognitive science involves the update strategy of global graph representations and reasoning on a personal cognitive graph by adopting the idea of graph neural networks and attention mechanisms. For the diagnosis of diseases in general practice, the CRKG outperforms all the baselines with a precision@1 of 0.7873, recall@10 of 0.9020 and hits@10 of 0.9340. Additionally, the visualization of the reasoning process for each visit of a patient based on the knowledge graph enhances clinicians' comprehension and contributes to explainability. This study is of great importance for the exploration and application of decision making based on EHRs and knowledge graph.

Theoretical limits on the speed of learning inverse models explain the rate of adaptation in arm reaching tasks

An essential aspect of human motor learning is the formation of inverse models, which map desired actions to motor commands. Inverse models can be learned by adjusting parameters in neural circuits to minimize errors in the performance of motor tasks through gradient descent. However, the theory of gradient descent establishes limits on the learning speed. Specifically, the eigenvalues of the Hessian of the error surface around a minimum determine the maximum speed of learning in a task. Here, we use this theoretical framework to analyze the speed of learning in different inverse model learning architectures in a set of isometric arm-reaching tasks. We show theoretically that, in these tasks, the error surface and, thus the speed of learning, are determined by the shapes of the force manipulability ellipsoid of the arm and the distribution of targets in the task. In particular, rounder manipulability ellipsoids generate a rounder error surface, allowing for faster learning of the inverse model. Rounder target distributions have a similar effect. We tested these predictions experimentally in a quasi-isometric reaching task with a visuomotor transformation. The experimental results were consistent with our theoretical predictions. Furthermore, our analysis accounts for the speed of learning in previous experiments with incompatible and compatible virtual surgery tasks, and with visuomotor rotation tasks with different numbers of targets. By identifying aspects of a task that influence the speed of learning, our results provide theoretical principles for the design of motor tasks that allow for faster learning.

Physical effects of learning

Interacting many-body physical systems ranging from neural networks in the brain to folding proteins to self-modifying electrical circuits can learn to perform diverse tasks. This learning, both in nature and in engineered systems, can occur through evolutionary selection or through dynamical rules that drive active learning from experience. Here, we show that learning in linear physical networks with weak input signals leaves architectural imprints on the Hessian of a physical system. Compared to a generic organization of the system components, (a) the effective physical dimension of the response to inputs decreases, (b) the response of physical degrees of freedom to random perturbations (or system "susceptibility") increases, and (c) the low-eigenvalue eigenvectors of the Hessian align with the task. Overall, these effects embody the typical scenario for learning processes in physical systems in the weak input regime, suggesting ways of discovering whether a physical network may have been trained.

Learning beyond sensations: How dreams organize neuronal representations

Semantic representations in higher sensory cortices form the basis for robust, yet flexible behavior. These representations are acquired over the course of development in an unsupervised fashion and continuously maintained over an organism's lifespan. Predictive processing theories propose that these representations emerge from predicting or reconstructing sensory inputs. However, brains are known to generate virtual experiences, such as during imagination and dreaming, that go beyond previously experienced inputs. Here, we suggest that virtual experiences may be just as relevant as actual sensory inputs in shaping cortical representations. In particular, we discuss two complementary learning principles that organize representations through the generation of virtual experiences. First, "adversarial dreaming" proposes that creative dreams support a cortical implementation of adversarial learning in which feedback and feedforward pathways engage in a productive game of trying to fool each other. Second, "contrastive dreaming" proposes that the invariance of neuronal representations to irrelevant factors of variation is acquired by trying to map similar virtual experiences together via a contrastive learning process. These principles are compatible with known cortical structure and dynamics and the phenomenology of sleep thus providing promising directions to explain cortical learning beyond the classical predictive processing paradigm.

Inferring neural activity before plasticity as a foundation for learning beyond backpropagation

For both humans and machines, the essence of learning is to pinpoint which components in its information processing pipeline are responsible for an error in its output, a challenge that is known as 'credit assignment'. It has long been assumed that credit assignment is best solved by backpropagation, which is also the foundation of modern machine learning. Here, we set out a fundamentally different principle on credit assignment called 'prospective configuration'. In prospective configuration, the network first infers the pattern of neural activity that should result from learning, and then the synaptic weights are modified to consolidate the change in neural activity. We demonstrate that this distinct mechanism, in contrast to backpropagation, (1) underlies learning in a well-established family of models of cortical circuits, (2) enables learning that is more efficient and effective in many contexts faced by biological organisms and (3) reproduces surprising patterns of neural activity and behavior observed in diverse human and rat learning experiments.

Time-varying generalized linear models: characterizing and decoding neuronal dynamics in higher visual areas

To create a behaviorally relevant representation of the visual world, neurons in higher visual areas exhibit dynamic response changes to account for the time-varying interactions between external (e.g., visual input) and internal (e.g., reward value) factors. The resulting high-dimensional representational space poses challenges for precisely quantifying individual factors' contributions to the representation and readout of sensory information during a behavior. The widely used point process generalized linear model (GLM) approach provides a powerful framework for a quantitative description of neuronal processing as a function of various sensory and non-sensory inputs (encoding) as well as linking particular response components to particular behaviors (decoding), at the level of single trials and individual neurons. However, most existing variations of GLMs assume the neural systems to be time-invariant, making them inadequate for modeling nonstationary characteristics of neuronal sensitivity in higher visual areas. In this review, we summarize some of the existing GLM variations, with a focus on time-varying extensions. We highlight their applications to understanding neural representations in higher visual areas and decoding transient neuronal sensitivity as well as linking physiology to behavior through manipulation of model components. This time-varying class of statistical models provide valuable insights into the neural basis of various visual behaviors in higher visual areas and hold significant potential for uncovering the fundamental computational principles that govern neuronal processing underlying various behaviors in different regions of the brain.

Population encoding of stimulus features along the visual hierarchy

The retina and primary visual cortex (V1) both exhibit diverse neural populations sensitive to diverse visual features. Yet it remains unclear how neural populations in each area partition stimulus space to span these features. One possibility is that neural populations are organized into discrete groups of neurons, with each group signaling a particular constellation of features. Alternatively, neurons could be continuously distributed across feature-encoding space. To distinguish these possibilities, we presented a battery of visual stimuli to the mouse retina and V1 while measuring neural responses with multi-electrode arrays. Using machine learning approaches, we developed a manifold embedding technique that captures how neural populations partition feature space and how visual responses correlate with physiological and anatomical properties of individual neurons. We show that retinal populations discretely encode features, while V1 populations provide a more continuous representation. Applying the same analysis approach to convolutional neural networks that model visual processing, we demonstrate that they partition features much more similarly to the retina, indicating they are more like big retinas than little brains.

Connectome-based reservoir computing with the conn2res toolbox

The connection patterns of neural circuits form a complex network. How signaling in these circuits manifests as complex cognition and adaptive behaviour remains the central question in neuroscience. Concomitant advances in connectomics and artificial intelligence open fundamentally new opportunities to understand how connection patterns shape computational capacity in biological brain networks. Reservoir computing is a versatile paradigm that uses high-dimensional, nonlinear dynamical systems to perform computations and approximate cognitive functions. Here we present conn2res: an open-source Python toolbox for implementing biological neural networks as artificial neural networks. conn2res is modular, allowing arbitrary network architecture and dynamics to be imposed. The toolbox allows researchers to input connectomes reconstructed using multiple techniques, from tract tracing to noninvasive diffusion imaging, and to impose multiple dynamical systems, from spiking neurons to memristive dynamics. The versatility of the conn2res toolbox allows us to ask new questions at the confluence of neuroscience and artificial intelligence. By reconceptualizing function as computation, conn2res sets the stage for a more mechanistic understanding of structure-function relationships in brain networks.

Unravelling individual rhythmic abilities using machine learning

Humans can easily extract the rhythm of a complex sound, like music, and move to its regular beat, like in dance. These abilities are modulated by musical training and vary significantly in untrained individuals. The causes of this variability are multidimensional and typically hard to grasp in single tasks. To date we lack a comprehensive model capturing the rhythmic fingerprints of both musicians and non-musicians. Here we harnessed machine learning to extract a parsimonious model of rhythmic abilities, based on behavioral testing (with perceptual and motor tasks) of individuals with and without formal musical training (n = 79). We demonstrate that variability in rhythmic abilities and their link with formal and informal music experience can be successfully captured by profiles including a minimal set of behavioral measures. These findings highlight that machine learning techniques can be employed successfully to distill profiles of rhythmic abilities, and ultimately shed light on individual variability and its relationship with both formal musical training and informal musical experiences.

Functional neuronal circuits emerge in the absence of developmental activity

The complex neuronal circuitry of the brain develops from limited information contained in the genome. After the genetic code instructs the birth of neurons, the emergence of brain regions, and the formation of axon tracts, it is believed that temporally structured spiking activity shapes circuits for behavior. Here, we challenge the learning-dominated assumption that spiking activity is required for circuit formation by quantifying its contribution to the development of visually-guided swimming in the larval zebrafish. We found that visual experience had no effect on the emergence of the optomotor response (OMR) in dark-reared zebrafish. We then raised animals while pharmacologically silencing action potentials with the sodium channel blocker tricaine. After washout of the anesthetic, fish could swim and performed with 75-90% accuracy in the OMR paradigm. Brain-wide imaging confirmed that neuronal circuits came 'online' fully tuned, without requiring activity-dependent plasticity. Thus, complex sensory-guided behaviors can emerge through activity-independent developmental mechanisms.

Visual looming is a primitive for human emotion

Looming objects afford threat of collision across the animal kingdom. Defensive responses to looming and neural computations for looming detection are strikingly conserved across species. In mammals, information about rapidly approaching threats is conveyed from the retina to the midbrain superior colliculus, where variables that indicate the position and velocity of approach are computed to enable defensive behavior. Although neuroscientific theories posit that midbrain representations contribute to emotion through connectivity with distributed brain systems, it remains unknown whether a computational system for looming detection can predict both defensive behavior and phenomenal experience in humans. Here, we show that a shallow convolutional neural network based on the Drosophila visual system predicts defensive blinking to looming objects in infants and superior colliculus responses to optical expansion in adults. Further, the responses of the convolutional network to a broad array of naturalistic video clips predict self-reported emotion largely on the basis of subjective arousal. Our findings illustrate how motor and experiential components of human emotion relate to species-general systems for survival in unpredictable environments.