Neural Decoding and "Inner" Psychophysics: A Distance-to-Bound Approach for Linking Mind, Brain, and Behavior

Area of Science:

• Cognitive Neuroscience
• Psychophysics
• Computational Neuroscience

Background:

• A central challenge in cognitive neuroscience is understanding the neural implementation of psychological constructs.
• Fechner's concept of "inner psychophysics" aimed to precisely map mental quantities to brain activity, an ambition yet to be fully realized.
• Multivariate pattern analysis (MVPA), also known as neural decoding, offers a novel framework for this endeavor.

Purpose of the Study:

• To propose and validate a novel approach to inner psychophysics using MVPA.
• To link neural activity patterns, decision-making models, and behavioral measures like reaction time.
• To establish a method for quantitatively relating brain states to psychological experience.

Main Methods:

• Utilizing simple linear classifiers applied to high-dimensional neural activation spaces.
• Developing an "neural distance-to-bound" approach based on the architecture of linear classifiers and signal detection theory.
• Predicting reaction times based on the estimated distance from decision boundaries in activation space.

Main Results:

• Demonstrated the empirical viability of the neural distance-to-bound approach in visual object recognition tasks.
• Showcased the potential of linear classifiers within MVPA to serve as a foundation for inner psychophysics.
• Established a quantitative link between neural representations and behavioral measures (reaction time).

Conclusions:

• The proposed neural distance-to-bound approach offers a practical and generalizable method for developing an inner psychophysics.
• MVPA and linear decoding provide powerful tools for bridging the gap between neural activity and psychological theory.
• This work represents a significant step towards a comprehensive understanding of the mind-brain-behavior relationship.