Quantifying human startle response magnitude: Effects of filter passband and integrator time constant on eyeblink EMG response peak and area

Several methods of quantifying eyeblink EMG magnitude in adult humans were compared. In Experiment 1, eyeblinks were elicited with either 100 dB(A) broadband acoustic stimuli (N = 18) or 9 milliamp electrical pulses to the forehead (N = 17). The peak magnitude and area of eyeblink EMG were each sampled on Five input lines: raw unfiltered, raw with a passband of either 1-1000 Hz or 90-250 Hz, and integrated with a time constant of either 10 or 100 ms. The measurement of response magnitude was generally reduced by narrowing the filter passband or by increasing the integrator time constant. Filter passband had a more pronounced effect on the area measure, whereas integration had a more pronounced effect on the peak measure. All methods of quantifying blink magnitude were significantly correlated with all other methods (within-subiects correlations, P <.01), both before and after response habituation. In Experiment 2 (N = 33), blink-eliciting stimuli were 85, 95, and 105 dB acoustic pulses, preceded on some trials by a 60 or 70 dB acoustic pulse at a 120 ms lead interval. In Experiment 3 (N = 28), blink-eliciting stimuli were electrical pulses to the forehead, preceded on some trials by a 70 dB acoustic pulse at a 120 ms lead interval, or followed by the 70 dB acoustic pulse at lead intervals of 0, 25, 50, 75, or 100 ms. In all stimulus conditions in both Experiments 2 and 3, the Four methods of eyeblink response magnitude quantification (peak magnitude and area of raw EMG, each with either a 90-250 Hz passband or integrated with a 10 ms time constant) were highly correlated (P < .01). These data suggest that if eyeblink magnitude is the only measure of interest, all of the quantification methods tested will provide similar information. This conclusion increases the confidence with which one can compare data From different laboratories that use different eyeblink magnitude quantification methods.