Decoding EEG Brain Activity for Multi-Modal Natural Language Processing

Detection of Language Lateralization Using Spectral Analysis of EEG

PURPOSE

Language lateralization relies on expensive equipment and can be difficult to tolerate. We assessed if lateralized brain responses to a language task can be detected with spectral analysis of electroencephalography (EEG).

METHODS

Twenty right-handed, neurotypical adults (28 ± 10 years; five males) performed a verb generation task and two control tasks (word listening and repetition). We measured changes in EEG activity elicited by tasks (the event-related spectral perturbation [ERSP]) in the theta, alpha, beta, and gamma frequency bands in two language (superior temporal and inferior frontal [ST and IF]) and one control (occipital [Occ]) region bilaterally. We tested whether language tasks elicited (1) changes in spectral power from baseline (significant ERSP) at any region or (2) asymmetric ERSPs between matched left and right regions.

RESULTS

Left IF beta power (-0.37±0.53, t = -3.12, P = 0.006) and gamma power in all regions decreased during verb generation. Asymmetric ERSPs (right > left) occurred between the (1) IF regions in the beta band (right vs. left difference of 0.23±0.37, t(19) = -2.80, P = 0.0114) and (2) ST regions in the alpha band (right vs. left difference of 0.48±0.63, t(19) = -3.36, P = 0.003). No changes from baseline or hemispheric asymmetries were noted in language regions during control tasks. On the individual level, 16 (80%) participants showed decreased left IF beta power from baseline, and 16 showed ST alpha asymmetry. Eighteen participants (90%) showed one of these two findings.

CONCLUSIONS

Spectral EEG analysis detects lateralized responses during language tasks in frontal and temporal regions. Spectral EEG analysis could be developed into a readily available language lateralization modality.

The ZuCo benchmark on cross-subject reading task classification with EEG and eye-tracking data

We present a new machine learning benchmark for reading task classification with the goal of advancing EEG and eye-tracking research at the intersection between computational language processing and cognitive neuroscience. The benchmark task consists of a cross-subject classification to distinguish between two reading paradigms: normal reading and task-specific reading. The data for the benchmark is based on the Zurich Cognitive Language Processing Corpus (ZuCo 2.0), which provides simultaneous eye-tracking and EEG signals from natural reading of English sentences. The training dataset is publicly available, and we present a newly recorded hidden testset. We provide multiple solid baseline methods for this task and discuss future improvements. We release our code and provide an easy-to-use interface to evaluate new approaches with an accompanying public leaderboard: www.zuco-benchmark.com.

Learning neural decoders without labels using multiple data streams

OBJECTIVE

Recent advances in neural decoding have accelerated the development of brain-computer interfaces aimed at assisting users with everyday tasks such as speaking, walking, and manipulating objects. However, current approaches for training neural decoders commonly require large quantities of labeled data, which can be laborious or infeasible to obtain in real-world settings. Alternatively, self-supervised models that share self-generated pseudo-labels between two data streams have shown exceptional performance on unlabeled audio and video data, but it remains unclear how well they extend to neural decoding.

APPROACH

We learn neural decoders without labels by leveraging multiple simultaneously recorded data streams, including neural, kinematic, and physiological signals. Specifically, we apply cross-modal, self-supervised deep clustering to train decoders that can classify movements from brain recordings. After training, we then isolate the decoders for each input data stream and compare the accuracy of decoders trained using cross-modal deep clustering against supervised and unimodal, self-supervised models.

MAIN RESULTS

We find that sharing pseudo-labels between two data streams during training substantially increases decoding performance compared to unimodal, self-supervised models, with accuracies approaching those of supervised decoders trained on labeled data. Next, we extend cross-modal decoder training to three or more modalities, achieving state-of-the-art neural decoding accuracy that matches or slightly exceeds the performance of supervised models. Significance: We demonstrate that cross-modal, self-supervised decoding can be applied to train neural decoders when few or no labels are available and extend the cross-modal framework to share information among three or more data streams, further improving self-supervised training.