Nara, S. 1 , Zarraga , A. . 1 , Bourguignon, M. . 1, 3 & Molinaro , N. . 1, 2
1 Basque Center on Cognition, Brain and Language (BCBL), Spain
2 Ikerbasque, Basque Foundation for Science, Spain
3 ULB Neurosciences Institute, Université libre de Bruxelles, Belgium
The neurophysiological oscillatory correlates supporting predictive processing in the visual domain have been largely investigated [Howhy 2013, Clark 2016]. However, it is not clear how this predictive processing is sensitive to time uncertainty. In the present study, we evaluate the magnetoencephalographic correlates of visual prediction depending on the timing of the target stimulus.
Participants were presented with four visual gabor patches in a serial order (entrainers) followed, after a time lag by a fifth one (target). The orientation of the gabor could either change in either a gradual (predictable orientation) or a random (unpredictable orientation) manner across entrainers up to the target. Also, the timing of the entrainers can be either fixed (predictable timing) or variable (unpredictable timing). We evaluated the differential MEG correlates between predictable and unpredictable orientation when the timing was either fixed
or variable.
Twenty-five (25) healthy participants took part in the study. Data were acquired using 306 channel MEG (Elekta, Neuromag). Data analyses were performed employing Fieldtrip after maxwell filtering. The time frequency estimation was performed using a Hanning taper within a range of 1 to 35 Hz(resolution of 1 hz) .
We observed significant beta (~20 Hz) negative (predictable > unpredictable) clusters in pre-stimulus window (-500 - 0 msec). This negative cluster was more anteriorly distributed for the predictable timing and more posterior for the unpredictable. We also observed similar increased theta (5-8 Hz) responses (unpredictable > predictable) in the post-stimulus window (100 - 300 msec ) for the both timing conditions. The increased attentional demands for the unpredictable timing can have recruited the parietal regions at a larger extent. While when the timing is fixed the generation of predictions involve more anterior brain regions. Source level analysis of MEG data also support the time frequency results.