Fig. 5.

Group-level source dipole analysis of MEG gradiometer data. (A) Group means of source dipole locations (shown as circles) overlaid onto an MNI-space template brain for the neural effect of prior knowledge (matching–mismatching) and perceptual learning {[post–pretest (6+24 channels)] < [post–pretest (one channel)]}. The spatial extent of the ellipses represents ± two SEMs. Also shown (yellow circle) is the location prior mean used for the dipole estimation procedure. (B) Dipole source strength in the left and right STG as a function of increasing sensory detail (24 channels vs. one channel), prior knowledge (matching–mismatching), and perceptual learning (post- vs. pretests). Error bars indicate ± two within-subject SEMs. Asterisks show the significance of planned comparisons: **P < 0.01; ***P < 0.001. (C) Right STG dipole within which the effect of prior knowledge (matching–mismatching conditions) on source strength correlated across subjects with the behavioral magnitude of perceptual learning (improvement in speech recognition post- vs. pretest for six-channel speech). (D) Right STG dipole within which the effect of perceptual learning (post- vs. pretest averaged over six- and 24-channel speech) on source strength correlated across subjects with the magnitude of learning-related change in behavior (improvement in speech recognition post- vs. pretest for six-channel speech). ch, channels; M, matching; MM, mismatching.