Detection and discrimination of modulation type at low modulation rates

For a sinusoidal carrier at a frequency of 1 kHz, amplitude (AM) or frequency (FM) modulated by a sinusoidal modulator at a rate of fmod = 2, 5 or 10 Hz, psychometric functions for the detection and the discrimination of modulation type were measured as a function of an appropriate modulation index (i.e. m or β). Stimuli were presented in quiet or with a band of noise chosen to mask the low- or high-frequency side of the excitation pattern produced by a modulated signal. In AM case d' markedly depended on the presence of a noise that masked the high-frequency part of the excitation pattern. For FM signals, on the other hand, when modulation rate was equal to 2Hz bands of noise did not influence d' values. Probability of identification of modulation type (AM or FM) was the highest for the smallest modulation rate and it was nearly equal to the probability of modulation detection. Presence of any of these two bands of noise did not effect the modulation identification. The results suggest that there are two mechanisms underlying the detection and the discrimination of modulation type. One of them is based entirely on the changes in the excitation pattern level and operates for a whole range of carrier frequencies and modulation rates (place mechanism). However, for a low modulation rate there is another mechanism responsible for the detection of frequency changes only. This mechanism provides additional information about frequency changes and brings about markedly higher delectability d' for the detection and discrimination of frequency changes. This mechanism is not based on the ability of the auditory system to compare a phase of the excitation pattern changes at different frequency areas. It seems that information about frequency changes at a low rate may be effectively coded in a time distribution of neural spikes.