Natural Language Processing in Large-Scale Neural Models for Medical Screenings.

Area of Science:

• Computational neuroscience
• Cognitive science
• Medical diagnostics

Background:

• Medical screenings for neurological, psychological, and language disorders rely on assessing language and speech processing.
• Analyzing cognitive and linguistic impairments requires understanding the integrity of the neural language system.

Purpose of the Study:

• To propose a large-scale neural model simulating human behavior in medical screenings.
• To connect microscopic neural deficits within the model to macroscopic behavioral deficits.

Main Methods:

• Development of a large-scale neural model using the Neural Engineering Framework and Semantic Pointer Architecture.
• Simulation of two medical screenings: word naming and cognitive ability assessment.
• Comparison of model behavior in physiological (healthy) and pathological (impaired) cases.

Main Results:

• The model successfully simulated normal behavior in the physiological case for both screenings.
• Simulations with neural deficits demonstrated deviations in behavior, including increased errors.
• Different neural dysfunction types resulted in distinct behavioral deficit patterns and severity.

Conclusions:

• The proposed neural model effectively simulates cognitive, language, and speech processing relevant to medical screenings.
• The model provides a framework for understanding how neural deficits manifest as observable behavioral impairments.
• This approach aids in diagnosing and understanding neurological and language disorders.