An extremely fast neural mechanism to detect emotional visual stimuli: A two-experiment study

Defining the brain mechanisms underlying initial emotional evaluation is a key but unexplored clue to understanding affective processing. Event-related potentials (ERPs), especially suited for investigating this issue, were recorded in two experiments (n = 36 and n = 35). We presented emotionally negative (spiders) and neutral (wheels) silhouettes homogenized regarding their visual parameters. In Experiment 1, stimuli appeared at fixation or in the periphery (200 trials per condition and location), the former eliciting a N40 (39 milliseconds) and a P80 (or C1: 80 milliseconds) component, and the latter only a P80. In Experiment 2, stimuli were presented only at fixation (500 trials per condition). Again, an N40 (45 milliseconds) was observed, followed by a P100 (or P1: 105 milliseconds). Analyses revealed significantly greater N40-C1P1 peak-to-peak amplitudes for spiders in both experiments, and ANCOVAs showed that these effects were not explained by C1P1 alone, but that processes underlying N40 significantly contributed. Source analyses pointed to V1 as an N40 focus (more clearly in Experiment 2). Sources for C1P1 included V1 (P80) and V2/LOC (P80 and P100). These results and their timing point to low-order structures (such as visual thalamic nuclei or superior colliculi) or the visual cortex itself, as candidates for initial evaluation structures.

Convolutional neural networks for vision neuroscience: significance, developments, and outstanding issues

Convolutional Neural Networks (CNN) are a class of machine learning models predominately used in computer vision tasks and can achieve human-like performance through learning from experience. Their striking similarities to the structural and functional principles of the primate visual system allow for comparisons between these artificial networks and their biological counterparts, enabling exploration of how visual functions and neural representations may emerge in the real brain from a limited set of computational principles. After considering the basic features of CNNs, we discuss the opportunities and challenges of endorsing CNNs as in silico models of the primate visual system. Specifically, we highlight several emerging notions about the anatomical and physiological properties of the visual system that still need to be systematically integrated into current CNN models. These tenets include the implementation of parallel processing pathways from the early stages of retinal input and the reconsideration of several assumptions concerning the serial progression of information flow. We suggest design choices and architectural constraints that could facilitate a closer alignment with biology provide causal evidence of the predictive link between the artificial and biological visual systems. Adopting this principled perspective could potentially lead to new research questions and applications of CNNs beyond modeling object recognition.

Sensing fear: fast and precise threat evaluation in human sensory cortex

Animal models of threat processing have evolved beyond the amygdala to incorporate a distributed neural network. In human research, evidence has intensified in recent years to challenge the canonical threat circuitry centered on the amygdala, urging revision of threat conceptualization. A strong surge of research into threat processing in the sensory cortex in the past decade has generated particularly useful insights to inform the reconceptualization. Here, synthesizing findings from both animal and human research, we highlight sensitive, specific, and adaptable threat representations in the sensory cortex, arising from experience-based sculpting of sensory coding networks. We thus propose that the human sensory cortex can drive smart (fast and precise) threat evaluation, producing threat-imbued sensory afferents to elicit network-wide threat responses.

Non-conscious processing of fear faces: a function of the implicit self-concept of anxiety

BACKGROUND

Trait anxiety refers to a stable tendency to experience fears and worries across many situations. High trait anxiety is a vulnerability factor for the development of psychopathologies. Self-reported trait anxiety appears to be associated with an automatic processing advantage for threat-related information. Self-report measures assess aspects of the explicit self-concept of anxiety. Indirect measures can tap into the implicit self-concept of anxiety.

METHODS

We examined automatic brain responsiveness to non-conscious threat as a function of trait anxiety using functional magnetic resonance imaging. Besides a self-report instrument, we administered the Implicit Association Test (IAT) to assess anxiety. We used a gender-decision paradigm presenting brief (17 ms) and backward-masked facial expressions depicting disgust and fear.

RESULTS

Explicit trait anxiety was not associated with brain responsiveness to non-conscious threat. However, a relation of the implicit self-concept of anxiety with masked fear processing in the thalamus, precentral gyrus, and lateral prefrontal cortex was observed.

CONCLUSIONS

We provide evidence that a measure of the implicit self-concept of anxiety is a valuable predictor of automatic neural responses to threat in cortical and subcortical areas. Hence, implicit anxiety measures could be a useful addition to explicit instruments. Our data support the notion that the thalamus may constitute an important neural substrate in biased non-conscious processing in anxiety.

A deep neural network model of the primate superior colliculus for emotion recognition

Although sensory processing is pivotal to nearly every theory of emotion, the evaluation of the visual input as ‘emotional’ (e.g. a smile as signalling happiness) has been traditionally assumed to take place in supramodal ‘limbic’ brain regions. Accordingly, subcortical structures of ancient evolutionary origin that receive direct input from the retina, such as the superior colliculus (SC), are traditionally conceptualized as passive relay centres. However, mounting evidence suggests that the SC is endowed with the necessary infrastructure and computational capabilities for the innate recognition and initial categorization of emotionally salient features from retinal information. Here, we built a neurobiologically inspired convolutional deep neural network (DNN) model that approximates physiological, anatomical and connectional properties of the retino-collicular circuit. This enabled us to characterize and isolate the initial computations and discriminations that the DNN model of the SC can perform on facial expressions, based uniquely on the information it directly receives from the virtual retina. Trained to discriminate facial expressions of basic emotions, our model matches human error patterns and above chance, yet suboptimal, classification accuracy analogous to that reported in patients with V1 damage, who rely on retino-collicular pathways for non-conscious vision of emotional attributes. When presented with gratings of different spatial frequencies and orientations never ‘seen’ before, the SC model exhibits spontaneous tuning to low spatial frequencies and reduced orientation discrimination, as can be expected from the prevalence of the magnocellular (M) over parvocellular (P) projections. Likewise, face manipulation that biases processing towards the M or P pathway affects expression recognition in the SC model accordingly, an effect that dovetails with variations of activity in the human SC purposely measured with ultra-high field functional magnetic resonance imaging. Lastly, the DNN generates saliency maps and extracts visual features, demonstrating that certain face parts, like the mouth or the eyes, provide higher discriminative information than other parts as a function of emotional expressions like happiness and sadness. The present findings support the contention that the SC possesses the necessary infrastructure to analyse the visual features that define facial emotional stimuli also without additional processing stages in the visual cortex or in ‘limbic’ areas.

This article is part of the theme issue ‘Cracking the laugh code: laughter through the lens of biology, psychology and neuroscience’.

Fast Unconscious Processing of Emotional Stimuli in Early Stages of the Visual Cortex

Several cortical and subcortical brain areas have been reported to be sensitive to the emotional content of subliminal stimuli. However, the timing of these activations remains unclear. Our scope was to detect the earliest cortical traces of emotional unconscious processing of visual stimuli by recording event-related potentials (ERPs) from 43 participants. Subliminal spiders (emotional) and wheels (neutral), sharing similar low-level visual parameters, were presented at two different locations (fixation and periphery). The differential (peak-to-peak) amplitude from CP1 (77 ms from stimulus onset) to C2 (100 ms), two early visual ERP components originated in V1/V2 according to source localization analyses, was analyzed via Bayesian and traditional frequentist analyses. Spiders elicited greater CP1–C2 amplitudes than wheels when presented at fixation. This fast effect of subliminal stimulation—not reported previously to the best of our knowledge—has implications in several debates: 1) The amygdala cannot be mediating these effects, 2) latency of other evaluative structures recently proposed, such as the visual thalamus, is compatible with these results, 3) the absence of peripheral stimuli effects points to a relevant role of the parvocellular visual system in unconscious processing.