Abstract
Brain state-dependent transcranial magnetic stimulation (TMS) requires real-time identification of cortical excitability states. Current approaches deliver TMS during brain states that correlate with motor cortex (M1) excitability at the group level. Here, we hypothesized that machine learning classifiers could successfully discriminate between high and low M1 excitability states in individual participants using information obtained from low-density electroencephalography (EEG) signals. To test this, we analyzed a publicly available dataset that delivered 600 single TMS pulses to the right M1 during EEG and electromyography (EMG) recordings in 20 healthy adults. Multivariate pattern classification was used to discriminate between brain states during which TMS evoked small and large motor-evoked potentials (MEPs). Results show that personalized classifiers successfully discriminated between low and high M1 excitability states in 80% of tested participants. MEPs elicited during classifier-predicted high excitability states were significantly larger than those elicited during classifier-predicted low excitability states in 90% of tested participants. Personalized classifiers did not generalize across participants. Overall, results show that individual participants exhibit unique brain activity patterns which predict low and high M1 excitability states and that these patterns can be efficiently captured using low-density EEG signals. Our findings suggest that deploying individualized classifiers during brain state-dependent TMS may enable fully personalized neuromodulation in the future.