On the other hand, ∼57% of electrodes in the temporal lobe have a large mean phase difference at t = 500 ms when the IPC values are at their peak. Therefore, the phase difference is likely to be small just after the stimulus appears, Vorinostat cell line and the number of electrodes with large phase differences increases while the image is showing (consistent with Figure 6D). Note that, while the data before t = 0 appear smooth and may give an idea of the overall trend, they are not statistically significant. These

analyses highlight the key differences in the phase of LFPs between temporal and frontal regions and provide a clear picture of how the responses develop by first aligning in phase and later developing different means. In addition, the largest phase

differences in the temporal lobe coincide with the maximum values of IPC. This is consistent with the idea that a high d  ′ value is a product of both an increase in IPC and a large mean phase difference this website ( Rizzuto et al., 2006). More detailed analyses reveal that, as one may expect, dphase’ increases with both increased phase coherence and with phase difference between correct and incorrect trials ( Figure S3). The LFP responses observed during the memory task could be generated by different mechanisms. Earlier, we noted that alignment of phases across trials could be caused by a “reset” of ongoing oscillations (Figure 2B, right). If this is the case, the oscillation should be present before the stimulus, there should be an increase in phase coherence caused by the stimulus, and there should be no these associated increase in amplitude (Shah et al., 2004). Alternatively, the increase in IPC could be caused by the presence of a stimulus-evoked response added to ongoing activity (Figure 2B, middle). Such a signal would cause a temporary increase in power at the frequency in question. In practice, these two mechanisms are difficult to differentiate. Note that the additive evoked response and the phase reset can produce the same average across trials and

the induced oscillation produced no mean response (Figure 2B). Thus, the average signal is not a reliable way to identify the underlying mechanism. Instead, the responses in each electrode can be characterized by the mean amplitude over all trials and the IPC. Note that the amplitude is acquired from the wavelet transform of individual trials of LFP data, so a group of trials can have an increase in mean amplitude, even if mismatched phases cause the mean of the raw LFP signals to be zero. This is the case for the induced oscillation: there is an increase in mean amplitude due to the stimulus, but there is no increase in IPC (Figure 7A, green). The evoked potential produces an increase in both mean amplitude and IPC (Figure 7A, blue), and the phase reset causes an increase in IPC with no associated increase in mean amplitude (Figure 7A, red).