Accession Number : AD1016230


Title :   Cascading Oscillators in Decoding Speech: Reflection of a Cortical Computation Principle


Descriptive Note : Technical Report,01 Sep 2011,31 Aug 2016


Corporate Author : Trustees Of Boston University Boston United States


Personal Author(s) : Ghitza,Oded


Full Text : https://apps.dtic.mil/dtic/tr/fulltext/u2/1016230.pdf


Report Date : 06 Sep 2016


Pagination or Media Count : 26


Abstract : Motivated by the possible role of brain rhythms in cortical function, we postulate a cortical computation principle by which decoding is performed within a time-varying window structure, synchronized with the input on multiple time scales. The windows are generated by a segmentation process, implemented by an array of cascaded oscillators. Correct segmentation is a critical prerequisite for correct decoding, and segmentation is correct as long as the oscillators successfully track the input rhythms. Syllabic segmentation utilizes flexible oscillators operating in the theta range (39Hz) by tracking the input syllabic rhythms, and prosodic segmentation is driven by flexible oscillators in the delta range(0.53 Hz), tracking prosodic rhythms. A model (TEMPO) was developed which is capable of explaining a variety of psychophysical and neuroimaging data difficult to explain by current models of speech perception, but emerging naturally from the architecture of the model. The key properties that enable such accountability are: (i) the capability of the oscillators to track and stay locked to the input rhythm, and (ii) the cascaded nature of the oscillators within the array.


Descriptors :   automated speech recognition , oscillators , perception , DECODING , memory , PHASE LOCKED SYSTEMS


Subject Categories : Voice Communications


Distribution Statement : APPROVED FOR PUBLIC RELEASE