Comparison of spinal myotatic reflexes in human adults investigated with cross-correlation and signal averaging methods

A cross-correlation method for recording spinal myotatic reflexes has been developed to meet the need for brief test periods in babies and children and subjects with central neurological pathology. In normal adult subjects the method has been validated by comparing excitatory and inhibitory reflexes obtained with cross-correlation with those obtained with conventional signal averaging. In the cross-correlation method a pseudo-random binary sequence of 64 brief tendon taps was delivered in <1.5 s, and in the averaging method 20-150 taps at one per second. The reflexes were expressed as unit impulse responses to enable direct, quantitative comparisons to be made. With cross-correlation the responses were slightly expanded in time, had lower peak amplitudes, and onset latencies advanced by 10 ms, the clock period of the pseudo-random binary sequence. The amplitude of biceps phasic stretch reflex increased with muscle contraction in a similar manner with both methods. In tests for stationarity the amplitude of biceps phasic stretch reflex varied <10% in the first six repeats of the pseudo-random binary sequence. The tap force required at threshold for cross-correlation was approximately half that for averaging, but with both methods the magnitude of biceps phasic stretch reflex varied linearly with tap force over the range of one to two times threshold. The validity of responses obtained with cross-correlation was assessed by a statistical procedure. In conclusion, the cross-correlation method is robust and gives similar results to those obtained with averaging.