Neural correlates of psychometric functions for auditory detection
Huanping Dai · The Journal of the Acoustical Society of America · 2006
The psychometric function in a 2IFC task, which describes the relation between the percentage of correct responses (pc) and the signal strength (x), can be expressed using the Gaussian probability function (φ) with a threshold (α) and a slope (β) parameter, pc=φ[(x/α)β/21/2]. For auditory detection, the psychometric functions exhibited a wide range of slopes in different conditions, such as tone in noise [β≊1, D. M. Green, M. J. McKey, and J. C. R. Licklider, J. Acoust. Soc. Am. 31, 1446–1452 (1959)] and tone in quiet [β≊2, C. S. Watson, J. R. Franks, and D. C. Hood, J. Acoust. Soc. Am. 52, 633–643 (1972)] by listeners with acoustic hearing, and detection by listeners with electrical hearing [βmax≊10, G. S. Donaldson, N. F. Viemeister, and D. A. Nelson, J. Acoust. Soc. Am. 101, 3706–3721 (1997)]. To understand the mechanisms underlying the change in the slope of the psychometric function, the performance of a neural detector engaged in the same tasks as the listeners was analyzed, and the neural correlates of the psychometric function, the neurometric function, was derived. Under the assumption that the detection was based on spike counts of the neuron(s) tuned to the signal frequency, the derived neurometric function showed slope values in good agreement with those obtained from the listeners.