Online internal speech decoding from single neurons in a human participant

Abstract

Speech brain-machine interfaces (BMI's) translate brain signals into words or audio outputs, enabling communication for people having lost their speech abilities due to diseases or injury. While important advances in vocalized, attempted, and mimed speech decoding have been achieved, results for internal speech decoding are sparse, and have yet to achieve high functionality. Notably, it is still unclear from which brain areas internal speech can be decoded. In this work, a tetraplegic participant with implanted microelectrode arrays located in the supramarginal gyrus (SMG) and primary somatosensory cortex (S1) performed internal and vocalized speech of six words and two pseudowords. We found robust internal speech decoding from SMG single neuron activity, achieving up to 91% classification accuracy during an online task (chance level 12.5%). Evidence of shared neural representations between internal speech, word reading, and vocalized speech processes were found. SMG represented words in different languages (English/ Spanish) as well as pseudowords, providing evidence for phonetic encoding. Furthermore, our decoder achieved high classification with multiple internal speech strategies (auditory imagination/ visual imagination). Activity in S1 was modulated by vocalized but not internal speech, suggesting no articulator movements of the vocal tract occurred during internal speech production. This works represents the first proof-of-concept for a high-performance internal speech BMI.

Learning to operate an imagined speech Brain-Computer Interface involves the spatial and frequency tuning of neural activity

“…Fifteen healthy participants (5 females, mean age 23.9 years, SD ±2.3, range [19][20][21][22][23][24][25][26][27][28][29] took part in this study which was approved by the local Ethics Committee (Commission cantonale d'éthique de la recherche, project 2022-00451) and was performed in accordance with the Declaration of Helsinki. All participants provided written informed consent and received financial compensation for their participation.…”

“…Although previous studies have characterized the neural correlates of imagined speech (Ikeda et al, 2014), most often in comparison with overt speech (Brumberg et al, 2016;Leuthardt et al, 2012;Martin et al, 2016;Pei et al, 2011;Proix et al, 2022;Soroush et al, 2022), only a handful of BCI studies have attempted to decode imagined speech in real time, with promising but often limited effectiveness (Angrick et al, 2021;Leuthardt et al, 2011;Sereshkeh et al, 2017;Wandelt et al, 2022). This is due to different challenges and limitations primarily pertaining to the weakness of imagined speech signals as compared to overt speech (Brumberg et al, 2016;Leuthardt et al, 2012;Martin et al, 2014;Pei et al, 2011;Proix et al, 2022), the difficulty to precisely identify the onset of speech imagery (Martin et al, 2018b), inter-individual differences in the ability to control the BCI (Marchesotti et al, 2016), and to the technique employed to record brain activity.…”

Continuous synthesis of artificial speech sounds from human cortical surface recordings during silent speech production

“…This study would suggest that articulation imagery led to stronger activations in the PostCG and the MFG, whereas hearing imagery did so in the STG. Also, pointing toward the idea that imagined speech might be better decoded from sensory areas, a recent study [60] found that a small set of words could be consistently classified from human intracortical recordings in the supramarginal gyrus, posterior to the PostCG and mainly thought to be involved in language processing.…”

Decoding and geometry of ten finger movements in human posterior parietal cortex and motor cortex

“…[5][6][7]27] extended continuous decoding to arm control, with [27] controlling the user's own muscles. Recent work has also decoded speech from sensorimotor cortex [28][29][30][31]. However, relatively few BMI studies have focused on hand control [32][33][34][35][36][37], and previous studies frequently combine the ring and little fingers or leave them out altogether.…”