Speaker
Description
According to the predictive coding framework, mammalian brains make sense of the world by constantly comparing internal predictions with sensory input. Mismatch negativity, which reflects a mismatch of input and expectations, is influenced by mental disorders such as schizophrenia. While alterations in network dynamics could elucidate underlying causes of disorders, studying human network activity is constrained by both ethical and structural factors, raising interest in potential model organisms.
Here, information network dynamics in publicly available marmoset ECoG data during a roving auditory oddball task were analyzed using spectral multivariate transfer entropy. Results demonstrate consistent pre-stimulus information transfer from the prefrontal to the superior temporal regions, primarily carried by alpha- and beta-band activity, indicative of top-down predictions of upcoming stimuli. Differences for standards and deviants may reflect implicit learning of statistical regularities. Together, results provide preliminary evidence for preparatory hierarchical predictions of auditory stimuli in the common marmoset.