Statistical Analysis

With the current experiment, we wanted to assess whether (nonlinear) multisensory integration is mandatory for a PPS effect. Therefore, we conducted 2 distinct analyses. The first analysis was conducted based on the classic approach to study PPS in nonhuman primates, in which PPS is defined as a multisensory modulation of tactile processing due to an external sensory stimulation, as a function of the distance of these stimuli from the body in space (see Introduction). Therefore, to identify PPS electrodes, we used a 2-step statistical approach. First, we first identified electrodes responding to multisensory AT stimuli (vs. baseline), and among those electrodes, we investigated which responded in a manner suggesting multisensory integration (AT vs. A + T; see below). Second, among the electrodes showing a multisensory integration effect, we characterized those that had a PPS effect—a multisensory response that is dependent on the distance of exteroceptive signals (e.g., auditory information) to the body (see below for more details; a similar approach has been previously used in iEEG studies, e.g., Quinn et al. 2014). A second analysis aimed at identifying whether electrodes not showing multisensory integration did show a PPS effect. First, we first identified electrodes responding to the AT stimuli (vs. baseline). Second, among the electrodes showing a response, we identified those that had a PPS effect. For each of the above mentioned steps, statistical significance within each electrode was assessed through (temporal) cluster-based permutation statistics (Maris and Oostenveld 2007) as implemented in the Fieldtrip toolbox (Oostenveld et al. 2011). The advantage of this test is that differences between conditions can be identified without prior assumptions about the temporal distribution of effects. Therefore, it is a data-driven approach. The cluster-level statistic was calculated as the maximum sum (maxsum) of the t-values within the cluster. Statistical significance at the cluster level was determined by computing a Monte Carlo estimate of the permutation distribution of cluster statistics, using 5000 resampling of the original data, yielding a distribution of cluster-level statistics under the null hypothesis that any differences between conditions are due to chance. Within a single electrode, a cluster was taken to be significant if it fell outside the 95% confidence interval of the permutation distribution for that electrode. The determination of significant temporal clusters was performed independently for each electrode. This method controlled for false alarms within an electrode across time points.