Paper accepted in IEEE Transactions on Neural Systems & Rehabilitation Engineering

Liang, J., Bezsudnova, Y., Kowalczyk, A., Sun, Y., & Jensen, O. (2026). Enhanced neural decoding with optically pumped magnetometer MEG using multivariate pattern analysis. IEEE Trans. Neural Syst. Rehabil. Eng.

OPM-MEG outperformed SQUID-MEG for multivariate decoding of pictures and words, with the clearest advantage for word decoding. Decoding performance increased with sensor number and plateaued at around 40 sensors, with OPM reaching strong performance with fewer sensors the SQUID-MEG. Spatial-frequency analysis showed that OPM benefited more from higher-order components, suggesting better access to fine-grained neural information.