Lateralization of alpha power

Time-frequency analysis was used to estimate alpha power during presentation of digit stimuli in order to examine alpha lateralization. Using the Fieldtrip toolbox, wavelet decomposition with 7 wave cycles was applied to single-trial epochs. Power was estimated across the frequency range of 2 to 30 Hz in 1 Hz intervals. Single-trial power values at each time point and frequency were then normalized by dividing power values by the average of the wide-band power during the whole epoch. After, single-trial power values were then averaged across the time range during which the digit stimuli were presented, since digit stimulation varied from trial to trial. For each individual, alpha power was taken at each individual’s peak value in the 7–14 Hz band after averaging across trials and sensors³⁹. However, after peak alpha identification, these values were averaged separately for “attend left” and “attend right” conditions.

The convention for measuring alpha power lateralization involves computing a normalized “index” based on differences between the cued direction of spatial attention^16,40,41. Two indices have been used. First, the Attentional Modulation Index (AMI; [attend right − attend left] / [attend right + attend left]) indicates hemispheric changes in alpha power based on condition, with the expectation that positive values reflect larger alpha power in left hemisphere and negative values reflect larger alpha power in right hemisphere. Several studies have shown that the auditory alpha AMI contrasts most strongly in parietal sensors, and based on our equidistant EEG montage layout^42,43, we selected two symmetrical parietal sensors that reflected the largest positive and negative AMI values pooled across both groups (see Fig. 3A).

Ipsilateral suppression via alpha power modulation. (A) Topographies of the AMI for both groups. Large black circles represent sensors selected for analysis. Note that due to our equidistant layout, channel positions become distorted in 2D space near the vertex of the head. 3D plots of dorsal and posterior views adjacent to 2D topographies for each group shows accurate spatial locations of the sensors chosen for analysis. (B) AMI comparison for left hemisphere (LH) and right hemisphere (RH) sensors for both groups. Individual lines connect hemispheres within each subject. Thick colored bars represent the condition mean. (C) Comparison of ALI values between groups. Thick colored bars represent the group mean. * = p < 0.05.; n.s. = not significant. 2D and 3D scalp images were created using Fieldtrip software and custom routines in MATLAB.

The second index is Attentional Lateralization Index (ALI; [ipsilateral − contralateral] / [ipsilateral + contralateral]), and summarizes the condition-based changes in alpha power, irrespective of left- and right-sided differences. The ALI is computed from the same sensors as the AMI, and is precited to be positive and larger if alpha power is higher on the side ipsilateral to the cued attentional direction. The ALI is the primary measure to compare neural suppression via alpha modulation between NH and BiCI groups.