Dynamics and topography of cortical activity during time estimation: MEG and MEG-EEG studies

As yet controversial, the neural bases of time perception in the sub-second range have been widely explored through electroencephalography (EEG) or cerebral blood flow imaging techniques. Substantial improvement might be obtained by using whole-head magnetoencephalography (MEG), which provides a unique combination of both high temporal (millisecond) and spatial (millimeter) resolutions. Our study involved a duration discrimination task performed by human subjects monitored by either MEG or simultaneous MEG-EEG recordings. Subjects had to discriminate a previously memorized "standard" duration (700ms) among randomly presented test-stimuli ranging from 490ms to 910ms. The signals we used were either auditory (440Hz tones) or visual (LED light) in order to separate sensory modality-dependent activities and supramodal time-specific processing. Surface magnetic fields parallel the time course of the Contingent Negative Variation (CNV) demonstrated by EEG studies. But, while CNV is mainly a frontocentral slow wave, the magnetic signal peaks on temporal, temporo-parietal and parieto-occipital sensors. The time course and the topography of cortical activity were determined by solving the so-called "inverse problem". Based on individual head-models calculated from the segmented anatomical magnetic resonance imaging slices of each subject, we performed distributed source analyses with Minimum-Norm regularization. By using MEG-EEG fusion techniques, better focalization of electrophysiological activity maps was achieved. Our results show sustained activities in sensory areas respective to the signal's modality. The dynamics of these sources reproduces the duration of the stimuli. High inter-individual variability was observed in the topography of cortical activations correlative to the time estimation encompassing central premotor, intra-parietal, temporo-parieto-occipital and prefrontal areas.