Predictions not commands: active inference in the motor system

Adams, Rick A.; Shipp, Stewart; Friston, Karl J.

doi:10.1007/s00429-012-0475-5

Cited by 638 publications

(612 citation statements)

References 235 publications

(299 reference statements)

Supporting

Mentioning

590

Contrasting

Order By: Relevance

“…Note, however, that corticomuscular coherence was maximal at 9 Hz and hence di↵ers from typical alpha activity but may be related to the 'mu' rhythm (Pfurtscheller and Lopes Da Silva, 1999). Current motor control theories suggest that an internal model is used in trajectory planning to predict the sensory consequences of motor commands, which is updated through sensory prediction errors (Kawato, 1999;Wolpert and Ghahramani, 2000;Shadmehr et al, 2010;Friston, 2010;Adams et al, 2013). Error correction and adaption is then only required when there is a mismatch between the observed sensory signals and those predicted by the model.…”

Section: Discussionmentioning

confidence: 99%

The reorganization of corticomuscular coherence during a transition between sensorimotor states

2014

View full text Add to dashboard Cite

Recent research suggests that neural oscillations in di↵erent frequency bands support distinct and sometimes parallel processing streams in neural circuits. Studies of the neural dynamics of human motor control have primarily focused on oscillations in the beta band (15 30 Hz). During sustained muscle contractions, corticomuscular coherence is mainly present in the beta band, while coherence in the alpha (8 12 Hz) and gamma (30 80 Hz) bands has not been consistently found. Here we test the hypothesis that the frequency of corticomuscular coherence changes during transitions between sensorimotor states. Corticomuscular coherence was investigated in twelve participants making rapid transitions in force output between two targets. Corticomuscular coherence was present in the beta band during sustained contractions but vanished before movement onset, being replaced by transient synchronization in the alpha and gamma bands during dynamic force output. Analysis of the phase spectra suggested a time delay from muscle to cortex for alpha-band coherence, by contrast to a time delay from cortex to muscle for gamma-band coherence, indicating a↵erent and e↵erent corticospinal interactions respectively. Moreover, alpha and gamma -band coherence revealed distinct spatial topologies, suggesting di↵erent generative mechanisms. Coherence in the alpha and gamma bands was almost exclusively confined to trials showing a movement overshoot, suggesting a functional role related to error correction. We interpret the dual-band synchronization in the alpha and gamma bands as parallel streams of corticospinal processing involved in parsing prediction errors and generating new motor predictions.

show abstract

Section: Discussionmentioning

confidence: 99%

The reorganization of corticomuscular coherence during a transition between sensorimotor states

2014

View full text Add to dashboard Cite

show abstract

“…Our results can also be interpreted using an active inference framework (20). Here, intentions are abstract predictions about likely outcomes, which are Bayes-optimally combined with sensory evidence about outcomes when this becomes available.…”

Section: Exp 1: Involuntary Movementmentioning

confidence: 98%

Extending experiences of voluntary action by association

Khalighinejad

Haggard

2016

Proc. Natl. Acad. Sci. U.S.A.

View full text Add to dashboard Cite

"Sense of agency" refers to the experience that links one's voluntary actions to their external outcomes. It remains unclear whether this ubiquitous experience is hardwired, arising from specific signals within the brain's motor systems, or rather depends on associative learning, through repeated cooccurrence of voluntary movements and their outcomes. To distinguish these two models, we asked participants to trigger a tone by a voluntary keypress action. The voluntary action was always associated with an involuntary movement of the other hand. We then tested whether the combination of the involuntary movement and tone alone might now suffice to produce a sense of agency, even when the voluntary action was omitted. Sense of agency was measured using an implicit marker based on time perception, namely a shift in the perceived time of the outcome toward the action that caused it. Across two experiments, repeatedly pairing an involuntary movement with a voluntary action induced key temporal features of agency, with the outcome now perceived as shifted toward the involuntary movement. This shift required involuntary movements to have been previously associated with voluntary actions. We show that some key aspects of agency may be transferred from voluntary actions to involuntary movements. An internal volitional signal is required for the primary acquisition of agency but, with repeated association, the involuntary movement in itself comes to produce some key temporal features of agency over the subsequent outcome. This finding may explain how humans can develop an enduring sense of agency in nonnatural cases, like brainmachine interfaces. I n a series of brilliant experiments, Roger Sperry switched the nerves for flexion of the rat hind leg with the nerves for extension. After that, whenever the bottom of the foot was injured, the rat extended the foot instead of flexing it. Rats never learned to lift up the paw, and "no adaptive functioning of the nervous system took place" (1). When the optic nerves of salamanders were cut, and the eyeball rotated 180°, salamanders saw upside down for the rest of their lives (2). These experiments suggested that key sensorimotor brain circuits are largely hardwired, and impervious to modification by experience."Sense of agency" refers to the capacity to control one's actions and, through them, the external world. Sense of agency is fundamental to instrumental and goal-directed actions, and forms the cornerstone of humans' astonishing capacity to change their physical and social environment (3). However, it remains unclear how the brain produces this distinctive and important subjective experience. Some recent results have linked the sense of agency to specific preparatory volitional signals in frontal (4) and parietal (5) areas, which then trigger voluntary motor commands passing through the "final common path" (6) of the primary motor cortex. Importantly, these "volitional signals" were generated well before the occurrence of both action and outcome, and were strongly correlated ...

show abstract

“…Note that action is prescribed by predictions of the next outcome and is not itself part of the inference process. This separation of action and perceptual inference or state estimation can be understood by associating action with peripheral reflexes in the motor system that fulfill top-down motor predictions about how we move (Feldman, 2009;Adams, Shipp, & Friston, 2013).…”

Section: Active Inference and Learningmentioning

confidence: 99%

Active Inference: A Process Theory

et al. 2017

View full text Add to dashboard Cite

This is the published version of the paper.This version of the publication may differ from the final published version. behavior prescribed by these dynamics has a degree of face validity, providing a formal explanation for reward seeking, context learning, and epistemic foraging. Technically, the fact that a gradient descent appears to be a valid description of neuronal activity means that variational free energy is a Lyapunov function for neuronal dynamics, which therefore conform to Hamilton's principle of least action. Permanent repository link

show abstract

Predictions not commands: active inference in the motor system

Cited by 638 publications

References 235 publications

The reorganization of corticomuscular coherence during a transition between sensorimotor states

The reorganization of corticomuscular coherence during a transition between sensorimotor states

Extending experiences of voluntary action by association

Active Inference: A Process Theory

Contact Info

Product

Resources

About