Abstract
The Sense of Agency (SoA) refers to the subjective experience of being the author of one's actions and their subsequent effects on the external environment. This experience is fundamentally built upon two pillars: subjective intention and result feedback. This research framework utilizes high-resolution Magnetoencephalography (MEG) to investigate the spatiotemporal markers of SoA within the human brain. By manipulating the attributes of intention and feedback, the authors aim to map the feedforward and feedback mechanisms within the fronto-parietal network. Establishing these neural markers is crucial for understanding human action generation and providing objective diagnostic standards for psychiatric conditions like schizophrenia, where the sense of self-agency is often disrupted.
Introduction: Decoding the Sense of Agency
While the feeling of being in control of our movements seems intuitive, the underlying neural processes are remarkably complex. The "Sense of Agency" (SoA) is what allows us to distinguish between a self-initiated action (like reaching for a cup) and a passive movement (like someone moving our arm). Current cognitive models, such as the "Comparator Model," posit that the brain generates a prediction of the sensory consequences of an action and compares it with the actual feedback. However, empirical evidence regarding the precise timing and location of these neural computations remains fragmented.
This paper addresses the critical need for a unified cognitive neural model. By identifying specific spatiotemporal fingerprints, researchers can better understand how the brain integrates "top-down" intentions with "bottom-up" sensory information. Furthermore, these markers could serve as vital biomarkers for clinical diagnosis in disorders characterized by impaired self-awareness.
Proposed Methodology
The research framework centers on the use of Magnetoencephalography (MEG), a technique that offers millisecond-level temporal resolution alongside high spatial accuracy. This allows for the tracking of rapid neural dynamics across the cortex. The proposed program is divided into two primary experimental lines:
- Manipulating Intention: This study compares active versus passive movements to isolate "feedforward" signals. By varying the presence and specificity of intentions, researchers aim to identify markers in frontal regions such as the pre-supplementary motor area (pre-SMA) and the dorsolateral prefrontal cortex (DLPFC).
- Manipulating Feedback: This study focuses on the sensory outcomes of actions, varying their causal strength, modality (e.g., auditory vs. visual), and emotional valence. This approach targets the "feedback" processing in the posterior parietal cortex (PPC), angular gyrus, and insula.
A key behavioral metric used is "Intentional Binding"—the phenomenon where the perceived time interval between a voluntary action and its effect is subjectively compressed. This serves as an implicit measure of agency, complemented by neural markers like the Correct-related Negativity (CRN), Stimulus-preceding Negativity (SPN), and Feedback-related Negativity (FRN).
The Fronto-Parietal Network
A central thesis of this work is that the Sense of Agency emerges from a dynamic, distributed network rather than a single localized "agency center." The authors identify a fronto-parietal hub as the primary substrate for this experience.
As illustrated in Figure 1, the frontal regions (red) are primarily associated with the "before" phase—planning, intention, and the generation of forward models. Specifically, the pre-SMA and DLPFC are involved in the selection and initiation of voluntary acts. Conversely, the parietal and insular regions (blue) are involved in the "after" phase—monitoring outcomes and comparing them against internal predictions. The PPC and angular gyrus act as comparators, while the Insula (Ins) may integrate these signals into a coherent subjective experience of the self.
Key Insights and Theoretical Synthesis
The proposed framework synthesizes several critical findings from existing literature:
- Spatiotemporal Fingerprints: Different stages of agency are marked by distinct neural signatures. For instance, the SPN reflects the brain's anticipation of a result, while the CRN correlates with subjective confidence in the action's execution. Post-outcome, the M100 (early sensory processing) and FRN (prediction error) indicate the degree of match between intention and result.
- Sensory Attenuation: A hallmark of agency is the brain's ability to "mute" sensory responses to self-generated stimuli. This explains why self-tickling is ineffective; the brain predicts the sensation and attenuates the resulting neural response (e.g., the N100 wave).
- Clinical Implications: The framework highlights the relevance of these markers to psychiatry. Patients with schizophrenia often exhibit abnormal "intentional binding" or disrupted connectivity between the prefrontal cortex and the angular gyrus, suggesting a failure in the neural mechanisms that distinguish self from other.
Limitations and Future Directions
While the proposed framework provides a robust roadmap, several challenges remain. First, there is often a discrepancy between explicit reports of agency ("I did that") and implicit measures like intentional binding. Future research must reconcile these different levels of conscious experience. Second, most current studies rely on simple laboratory tasks (e.g., button presses). Increasing the "ecological validity" by studying more complex, real-world actions is essential. Finally, the ultimate goal is to translate these MEG markers into reliable clinical biomarkers, providing objective tools for the diagnosis and monitoring of psychiatric disorders involving the self.
Summary updated