The functional anatomy of attention: a DCM study

Area of Science:

• Neuroscience
• Cognitive Science
• Computational Neuroscience

Background:

• Attention is theorized to enhance sensory information processing by increasing its precision.
• Predictive coding frameworks link attentional gain to the expected precision of sensory inputs.
• The Posner paradigm provides a model for understanding attentional effects on target detection.

Purpose of the Study:

• To investigate the role of cue-dependent and top-down modulation of neural gain in attention.
• To test the predictions of a normative model of attention based on predictive coding using empirical data.
• To determine the neural sources and characteristics of attentional gain modulation.

Main Methods:

• Utilized dynamic causal modeling (DCM) to analyze magnetoencephalography (MEG) data.
• Applied Bayesian model comparison to assess the evidence for specific neural mechanisms.
• Focused on modeling superficial pyramidal cell gain as a proxy for prediction error signaling.

Main Results:

• Strong evidence supports the contribution of superficial pyramidal cell gain and its top-down modulation to observed neural responses.
• High certainty (>80%) that anticipatory effects on post-synaptic gain are localized to visual extrastriate sources.
• Findings align with predictive coding principles where attention optimizes inference through precision modulation.

Conclusions:

• Attention's role in optimizing perceptual inference is supported by empirical evidence.
• Top-down modulation of neural gain, particularly in visual areas, is a key mechanism.
• The study validates predictive coding as a framework for understanding attention and perception.