Predictive models avoid excessive reductionism in cognitive neuroimaging.

Area of Science:

• Cognitive neuroscience
• Computational neuroscience
• Neuroscience

Background:

• Understanding complex behavior requires integrated models of behavior, mental processes, and neural activity.
• Traditional cognitive neuroscience experiments often use reductionist approaches, limiting theory generalization.
• Paradigm-bound theories struggle to capture the full scope of brain-mind associations.

Purpose of the Study:

• To advocate for predictive modeling in cognitive neuroscience.
• To demonstrate how predictive models can generalize brain-mind associations to novel tasks and stimuli.
• To highlight the utility of predictive models for robust reverse inference and theory broadening.

Main Methods:

• Developing predictive models that link neural activity to behavior.
• Utilizing computational approaches to generalize findings across diverse tasks.
• Employing reverse inference techniques to identify neural correlates of mental processes.

Main Results:

• Predictive models successfully generalize brain-mind associations to new tasks and stimuli.
• Predicting behavior from neural activity enables robust reverse inference.
• Predicting neural activity from task descriptions allows for comprehensive modeling.

Conclusions:

• Predictive modeling is crucial for advancing cognitive neuroscience beyond paradigm-specific limitations.
• These models facilitate a deeper understanding of brain-mind relationships.
• Broader theories of cognition can be developed using predictive modeling frameworks.