A Neuronal Network Model for Human Echolocation.

A Neuronal Network Model for Human Echolocation.

A Neuronal Network Model for Human Echolocation.

Accession number;00A0101834

Title;A Neuronal Network Model for Human Echolocation.

Author;HOSHINO OSAMU(Oita Univ., Fac. of Eng.) TAHARA SHUNJI(Oita Univ., Fac. of Eng.) KUROIWA KAZUHARU(Oita Univ., Fac. of Eng.)

Journal Title;IEIC Technical Report (Institute of Electronics, Information and Communication Engineers)

Journal Code:S0532B

ISSN:0913-5685

VOL.99;NO.434(EA99 65-72);PAGE.23-30(1999)

Figure&Table&Reference;FIG.7, REF.7

Pub. Country;Japan

Language;Japanese

Abstract;We propose a neural network model of the inferior colliculus of the midbrain for echolocation in humans. The neurons of the network act as cross-correlators that receive contralateral inputs with time delays and ipsilateral inputs with no delays. We apply pairs of direct and echo sounds as artificial sonar sounds and their reflections from target objects, respectively. We train the network and modify the synaptic connection strengths according to the Hebbian rule. We demonstrate that a latency axis of the inferior colliculus, ranging from a few to tens of msec, is utilized for echolocation in humans. We also demonstrate that performance on echolocation is enhanced by self-organized synaptic depression in inhibitory synapses through the training processes. (author abst.)