Fig. 3. Statistical analysis description. (  A ) A difference phasor (  thin black vector ) is calculated for each frequency component (  e.g. , 40-Hz) by subtracting the recorded (  thick gray ) and predicted (  black )phasors containing the origin. The difference phasor represents the difference between the predicted and recorded signals in terms of amplitude and phase. The phasor analysis presented here for illustration pertains to data from an individual recording experiment from the current study. (  B )The 40-Hz difference phasor from  A (  circumscribed asterisk in the left lower quadrant ) is plotted with the difference phasors (  asterisks )from all recording experiments. The Hotelling  T ²test is used to determine whether the average difference phasor (  single vector containing the origin ) is significantly different from zero. Ellipses centered around the phasor tip are calculated that correspond to specific  P values. A difference phasor is statistically different from the origin if the ellipse does not contain the origin. The conclusion for this particular example is that the recorded and predicted phasors (and the corresponding time-domain waveforms) are statistically different from each other (  P < 0.01). Note that ellipse size is inversely proportional to  P value. Here for example, for  P = 0.001, the corresponding ellipse contains the origin. ASSR = auditory steady state response.