Search

Advancements in deep learning algorithms over the past decade have led to extensive developments in brain–computer interfaces (BCI). A promising imaging modality for BCI is magnetoencephalography (MEG), which is a non-invasive functional imaging technique. The present study developed a MEG sensor-based BCI neural network to decode Rock-Paper-scissors gestures (MEG-RPSnet). Unique preprocessing pipelines in tandem with convolutional neural network deep-learning models accurately classified gestures. On a single-trial basis, we found an average of 85.56% classification accuracy in 12 subjects. Our MEG-RPSnet model outperformed two state-of-the-art neural network architectures for electroencephalogram-based BCI as well as a traditional machine learning method, and demonstrated equivalent and/or better performance than machine learning methods that have employed invasive, electrocorticography-based BCI using the same task. In addition, MEG-RPSnet classification performance using an intra-subject approach outperformed a model that used a cross-subject approach. Remarkably, we also found that when using only central-parietal-occipital regional sensors or occipitotemporal regional sensors, the deep learning model achieved classification performances that were similar to the whole-brain sensor model. The MEG-RSPnet model also distinguished neuronal features of individual hand gestures with very good accuracy. Altogether, these results show that noninvasive MEG-based BCI applications hold promise for future BCI developments in hand-gesture decoding.

Additional file 1: Table S1. Differentially methylated positions (DMPs) in MRS, Army STARRS and PRISMO with and without corrections for smoking status and alcohol use. Table S2. SNPs within 500 bps upstream or downstream of the significant DMPs. Table S3. Differentially methylated positions (DMPs) in MRS with and without correction for main associated SNPs. Table S4. Correlations between blood and brain methylation levels for the top CpG sites based on external data. Table S5. Number of probes and samples removed at each stage of the quality control pipeline. Figure S1. Methylation values (B values) at cg05656210 for each cohort separately. Figure S2. Methylation values (B values) at cg05901543 for each cohort separately. Figure S3. Methylation values (B values) at cg16956686 for each cohort separately. Figure S4. Methylation values (B values) at cg18917957 for each cohort separately. Figure S5. Methylation values (B values) at cg12169700 for each cohort separately. Figure S6. Methylation values (B values) at cg20756026 for each cohort separately. Figure S7. Differentially methylated region at HLA-DBP1: 33048416-33048814. Figure S8. Differentially methylated region at HLA-DPB1: 33043976-33054001. Figure S9. Differentially methylated region at HLA-DRB1: 32547019-32557404. Figure S10. Differentially methylated region at HLA-DRB1: 32551851-32552331. Figure S11. Differentially methylated region at HOXA4: 27169572-27170638. Figure S12. Differentially methylated region at HOXA4: 27169740-27171528. Figure S13. Differentially methylated region at KCNE1: 35827824-35884508. Figure S14. Differentially methylated region at KCNE1: 35831697-35832365. Figure S15. Differentially methylated region at MAD1L1: 1885033-1885402. Figure S16. Differentially methylated region at NTRK1: 156814881-156815792. Figure S17. Differentially methylated region at SLC17A3: 25882327-25882560. Figure S18. Differentially methylated region at TRMT12: 125461772-125464547. Figure S19. Example of blood-brain correlations of methylation levels in cg05656210. PFC: prefrontal cortex, EC: entorhinal cortex, STG: superior temporal gyrus, CER: cerebellum.

SNP-exposure (WCadj associations) and SNP-outcome (PTSD associations) coefficients used in the MR analysis. The lines (see legend) represents the results observed considering different methods. (DOCX 383 kb)

Distribution of AFB PRS in PTSD cases and controls. (DOCX 361 kb)

Gene Ontology (GO) enrichments (Q valueâ

Distribution of WCadj PRS in PTSD cases and controls. (DOCX 333 kb)

Genetic correlation of WCadj PRS (SNP Nâ =â 31, PT =10â 8) with the other anthropometric traits associated with PTSD in women. (DOCX 12 kb)

Correlation between WCadj PRS and PTSD in men. (DOCX 12 kb)

PTSD genetic correlations tested with LD score regression analysis. (DOCX 12 kb)

Similarity network based on GO enrichment results related to the AFB-PTSD correlation. Bubble color indicates the GO P value; bubble size indicates the frequency of the GO term in the Gene Ontology Annotation (UniProt-GOA) Database which annotates all UniProt entries with GO terms. Highly similar GO terms are linked by edges in the graph, where the line width indicates the degree of similarity. (DOCX 179 kb)

Correlation between AFB PRS and PTSD in men. (DOCX 12 kb)