To view an abstract, select an author from the vertical list on the left.
2012 Grants - Pakhomov
Automated Semantic Indices for Early Detection of Cognitive Changes
Serguei Pakhomov, Ph.D.
Regents of the University of Minnesota—Twin Cities
2012 Development of New Cognitive Functional Instrumentation in Alzheimer's Disease
Existing tests of verbal fluency are widely used in clinical diagnosis of Alzheimer's disease. These tests are simple to administer; however, the standard scoring methods do not reliably detect changes in the very early predementia stages.
Serguei Pakhomov, Ph.D., proposes to develop a new method using automatic speech recognition and language technologies to analyze word timing and semantic information (how meaning is conveyed through signs and language). They plan to develop a specialized speech recognition system that will convert the voice of the person with Alzheimer's disease to a verbatim transcription. This tool will measure timing of words with subsecond precision and enable the analysis of hesitation, intonation and voice pitch informative of the individual's cognitive and emotional state. The proposed instrument will be implemented on mobile and telephone-based platforms, enabling remote voice collection useful for long-term in-home observation and large-scale clinical trials. The team will use an existing dataset from the Mayo Clinic Study of Aging and a cross-sectional sample of 90 volunteers with Alzheimer's disease, MCI and controls from the University of Minnesota Memory Clinic. The researchers will validate the new instrument on these datasets and relate instrument performance on semantic relatedness measures to neural markers of neurodegeneration obtained with fMRI.
This computerized cognitive evaluation methodology may prove to be a reliable way to characterize subtle early cognitive changes associated with risk of progressing to Alzheimer's disease.