Computer-brain interaction to optimize auditory stimuli based on neuronal responses

The authors have developed a self-optimizing way of finding the best excitatory frequencies (BEF) of auditory neurons. The ALOPEX-(algorithm of pattern extraction) algorithm uses feedback from the extracellular responses recorded from frog (Rana Pipiens) auditory neurons to optimize sound stimuli. This computer based system works in real time to iteratively find the neuron´s BEF. The stimulus is produced by controlling the Yamaha (YMF262) FM-chip on board a Sound Blaster 16 card. The response and stimulus are captured by a Dataq (DI-300F) high speed analog and digital I/O board. Programmable threshold logic is used to determine the number of action potentials in response to sinusoidal sound stimuli. The normalized tone ranges in frequency from 80 to 3150 Hz. The duration and intra-stimulus intervals are 270 and 500 msec respectively. The system dynamics and reliability were tested with the response calculated as a parabolic function of frequency. Actual results, using 10 stimulus presentations per iteration, showed that the setup is able to converge within 100 iterations. A companion demonstration that uses the parabolic equation to simulate response has been developed and will be presented as a demo