Boom, Max van den

Abstract Studying mammalian brain function aids our understanding of human brain evolution. We implanted a beagle with a prototype human neuromodulation platform that measures activity from the brain surface. One year later, a set of simple sensory tasks was performed, finding visual and somatosensory representation in the canine homologs of the expected areas in humans. Surprisingly, the canine anterior ectosylvian gyrus, which is anatomically homologous to human receptive speech areas, was selectively active during independent social reinforcement tasks. This suggests that human speech understanding may have evolved from more general mammalian brain structures that are specialized for social reinforcement.

Abstract Several studies have shown that mouth movements related to the pronunciation of individual phonemes are represented in the sensorimotor cortex. This would theoretically allow for brain computer interfaces that are capable of decoding continuous speech by training classifiers based on the activity in the sensorimotor cortex related to the production of individual phonemes. To address this, we investigated the decodability of trials with individual and paired phonemes (pronounced consecutively with one second interval) using activity in the sensorimotor cortex. Fifteen participants pronounced 3 different phonemes and 3 combinations of two of the same phonemes in a 7T functional MRI experiment. We confirmed that support vector machine (SVM) classification of single and paired phonemes was possible. Importantly, by combining classifiers trained on single phonemes, we were able to classify paired phonemes with an accuracy of 53% (33% chance level), demonstrating that activity of isolated phonemes is present and distinguishable in combined phonemes. A SVM searchlight analysis showed that the phoneme representations are widely distributed in the ventral sensorimotor cortex. These findings provide insights about the neural representations of single and paired phonemes. Furthermore, it supports the notion that speech BCI may be feasible based on machine learning algorithms trained on individual phonemes using intracranial electrode grids.

AbstractFunctional magnetic resonance imaging (fMRI) studies have often recorded robust univariate group effects in the amygdala of subjects exposed to emotional stimuli. Yet it is unclear to what extent this effect also holds true when multi‐voxel pattern analysis (MVPA) is applied at the level of the individual participant. Here we sought to answer this question. To this end, we combined fMRI data from two prior studies (N = 112). For each participant, a linear support vector machine was trained to decode the valence of emotional pictures (negative, neutral, positive) based on brain activity patterns in either the amygdala (primary region‐of‐interest analysis) or the whole‐brain (secondary exploratory analysis). The accuracy score of the amygdala‐based pattern classifications was statistically significant for only a handful of participants (4.5%) with a mean and standard deviation of 37% ± 5% across all subjects (range: 28–58%; chance‐level: 33%). In contrast, the accuracy score of the whole‐brain pattern classifications was statistically significant in roughly half of the participants (50.9%), and had an across‐subjects mean and standard deviation of 49% ± 6% (range: 33–62%). The current results suggest that the information conveyed by the emotional pictures was encoded by spatially distributed parts of the brain, rather than by the amygdala alone, and may be of particular relevance to studies that seek to target the amygdala in the treatment of emotion regulation problems, for example via real‐time fMRI neurofeedback training.