Effect of initial setting on convergence to optimal hearing aid setting using a simplex method

We examined the final preferred frequency gain response (i.e. final setting) selected using different starting points (i.e. initial settings) with a modified simplex procedure. These starting points included a frequency response close to the one recommended by NAL-R (Byrne and Dillon, 1986), a frequency response with the high and low frequency gain set to the maximum (max), a frequency response with the high and low frequency gain set to the minimum (min), and one in-between these values (mid). The time taken to converge to the final setting was also recorded. Subjects judged the clarity of short discourse passages presented at 63 dB(A) in the presence of babble noise (S/N = 0). The results indicated that the majority of subjects selected the same preferred frequency response regardless of initial settings. Variations in initial settings mainly affected the convergence time to the final choice. The frequency response recommended by NAL-R required the least time to converge, while those of the extremes (min and max) required the longest convergence time. Clinical implications are discussed.