The Electrophysiological Correlates of Processing Self- and Externally Generated Sensations
In order to interact effectively with the world, the brain must distinguish between sensations which are caused by one’s own actions and external sensations that are caused by the environment. This may be achieved through an efference copy-based forward model mechanism in which a copy of the motor c...
Saved in:
| Main Author: | |
|---|---|
| Contributors: | |
| Format: | Doctoral Thesis |
| Language: | English |
| Published: |
Philipps-Universität Marburg
2024
|
| Subjects: | |
| Online Access: | PDF Full Text |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| Summary: | In order to interact effectively with the world, the brain must distinguish between sensations which are caused by one’s own actions and external sensations that are caused by the environment. This may be achieved through an efference copy-based forward model mechanism in which a copy of the motor command is sent to the relevant sensory cortices with a prediction for the action’s sensory outcome. Studies show that neural and behavioural responses to self-initiated sensations are modulated relative to identical but externally presented stimuli. In EEG studies, self-generated auditory stimuli elicit smaller early ERP components (N1 and P2) than identical but externally generated stimuli. Furthermore, pre-movement neural activity (RP and LRP) may encode the action’s sensory consequences. Previous EEG studies have focussed largely on the auditory domain, with inconsistent results coming from experiments with visual stimuli. These studies also lacked optimal control conditions with passive movements and behavioural measures of perception.
In this dissertation, three studies with healthy subjects (two published and one submitted to a journal) were carried out to examine questions related to the neural and behavioural correlates of the forward model mechanism. Study I examined the sensory ERPs elicited by visual stimuli that were triggered by active (self-initiated) and passive (involuntary; finger moved by device) movements. Stimulus perception was measured using an intensity judgement task. Visual N1 and P2 ERP amplitudes were reduced in the active condition, indicating suppression of the self-initiated sensory input. There was no effect of movement in the behavioural task. However, suppression of the P2 component was correlated with behavioural suppression. This component might reflect higher-level processes such as conscious evaluation of perceived intensity.
In Study II, it was investigated whether RP and LRP encode information related to anticipating action-effect contingency and stimulus modality. When the action was followed immediately by a stimulus, RP differed between active and passive movements approximately 200 ms before the button press. This difference was not there when the sensory consequences were delayed by one second. Conversely, LRP encoded the movement type but not the action-effect contingency. This demonstrates a dissociation between RP and LRP, with RP representing higher-level processes, including anticipation of upcoming stimuli and LRP being related to low-level preparation for the movement execution. MVPA was also used to investigate whether action-effect prediction was represented across the whole scalp. Movement decoding (active vs passive) showed ramping accuracy in all conditions from around -800 ms onwards up to an accuracy of ∼ 85% at the movement. Accuracy was lower in the control than in the visual and auditory conditions approximately 200 ms before the movement. Sensory modality (visual vs auditory) was also decodable for both active and passive conditions. The active condition showed increased accuracy shortly before the movement. The results suggest that pre-movement EEG activity encodes action-feedback prediction.
In Study III, in addition to active (executed here with a minimum latency of 700 ms) and passive movements, participants made quick movements (as quickly as possible in response to a tone cue). The active and quick conditions showed reduced N1-P2 amplitudes relative to the passive condition and did not significantly differ from each other. Active and quick also showed comparable LRP amplitudes, which were significantly greater than passive. However, though all three conditions showed a negative shift in RP, the quick condition had lower amplitudes than the active condition, indicating differences in motor preparation. This may be related to additional preparatory processes for executing the movement as quickly as possible after the cue. The results showed that, even though active and quick were prepared differently, this did not ultimately lead to differences in feedback processing.
Taken together, this thesis offers a finer-grained specification of the efference copy mechanism. There was robust evidence of neural sensory suppression in the visual domain. For the first time, it was shown that action-feedback processing is similar between active and quick movements that nonetheless differ in movement initiation, intention to move, task demand and preparation time. The studies have also demonstrated the dissociation between RP and LRP, with the novel result that LRP is significantly reduced preceding involuntary movements. Furthermore, across three studies, there was evidence that RP encodes higher-level motor preparation processes, including feedback anticipation, which was specific to the active condition. The thesis also contributed an innovative analysis approach using MVPA, which demonstrated prediction for the action’s outcome before the movement, taking into account patterns of activity across the entire scalp.
The studies presented in this thesis enhance our understanding of the efference copy mechanism, with potential implications for future translational work which could contribute to understanding the deficits associated with major symptoms of psychosis. |
|---|---|
| DOI: | 10.17192/z2024.0320 |