The analysis of MEG data commonly involves arithmetic averaging of the evoked magnetic fi elds (AEFs) across subjects. A tacit assumption of this practise is that signals from different subjects simply differ by some amount. If true, the standard deviation (SD) of signal strengths across subjects should be independent of the mean signal amplitude. We exploit the dependence of the M100- peak amplitude on the stimulus onset interval (SOI) to scrutinize the variation of SD with mean amplitude. Fifteen subjects were stimulated monaurally with tones of SOIs varying between 250 ms and 12.5 s. For each subject and hemisphere, the M100-peak amplitude was determined, either from the channel with the largest signal or from a cluster of fi ve channels whose signals were averaged arithmetically or geometrically. Irrespective of the measure, we found that the SD of the M100-peak amplitude across subjects scales linearly with the amplitude of the corresponding arithmetic mean, but remains practically constant for the amplitude of the geometric mean. This shows that the M100 amplitudes of different subjects, hemispheres, and stimulus conditions differ by factors rather than by amounts. We also show that the geometric mean of the M100 amplitudes across different conditions provides an excellent measure for normalizing such data. Our results question the common practise of arithmetic averaging across subjects, or subtraction procedures between conditions, and thus are likely to have widespread implications.
JavaScript jest wyłączony w Twojej przeglądarce internetowej. Włącz go, a następnie odśwież stronę, aby móc w pełni z niej korzystać.