Karl John Friston

The paper shows that metastable oscillatory modes in the brain emerge from synchronization in the spacetime connectome, matching MEG data from healthy individuals.

The paper investigates how burst characteristics in beta rhythms of the motor cortex differ across movement states and how they contribute to state classification and cortical processing.

This paper introduces a method to estimate anisotropy from neural timeseries using a generalized Lagrangian model and applies it to murine calcium imaging data.

This paper introduces a generative model of auditory word repetition that simulates how words are produced and recognized, showing it can infer and repeat target words from acoustic signals.

This paper shows canonical neural networks perform active inference by minimizing variational free energy through delayed plasticity and threshold adaptation.

The paper investigates altered effective connectivity in sensorimotor cortices using resting-state fMRI and DCM, finding changes associated with depression severity and treatment response.

The paper shows that musicianship and melodic predictability increase neural gain in auditory cortex during pitch deviance detection, as revealed by dynamic causal modeling of MEG data.

This paper uses dynamic causal models to infer muscarinic receptor function from EEG data in rats, showing graded effects and predictive validity across pharmacological conditions.

The paper investigates how hand-object orientation congruency affects neural activity during action planning using fMRI, finding increased premotor cortex activity for incongruent orientations.

This study shows that motor skill learning involves coherent waves of myelin plasticity in the corticospinal-hippocampal loop, with spatial and temporal patterns linked to training performance.

The paper simulates functional recovery from different lesion types using a generative model, showing that complete disconnection of a system invokes the alternative system, while partial damage has more severe long-term effects.

This paper develops a hierarchical active inference model for visual scene construction, showing it generates empirical evidence accumulation phenomena and novel behaviors in uncertain environments.

This study shows that multiple sclerosis alters higher-order BOLD responses in BA 4p but not BA 4a during grip force tasks, beyond the main effect of movement.

The paper investigates the neural mechanisms linking speech-in-noise perception difficulties to fundamental auditory grouping using fMRI, finding shared cortical processes in early auditory areas.

The paper uses Active Inference to simulate agents learning habit formation strategies in different environments, finding that specialization is advantageous in stable environments while generalization occurs in volatile ones.

The paper investigates how cortical beta oscillations reflect contextual gating of visual action feedback during sensorimotor integration using MEG and virtual reality.

This paper introduces active listening as a framework for speech synthesis and recognition, showing how word boundaries are selected to maximise evidence for internal models.

This paper introduces an active inference framework for modeling structure learning, using concept learning as an example, and demonstrates its ability to generalize and reduce model complexity.

The paper investigates how sedation with propofol affects predictive coding circuits and the double surprise acceleration effect using EEG source localization in a local-global task.

The paper investigates neural differences in metacognition for detecting versus discriminating visual stimuli using fMRI, finding distinct confidence effects in frontopolar cortex and right temporoparietal junction.

This study investigates how attention modulates task performance under conflicting visual and proprioceptive cues using simulation and human experiments.

This paper presents an Active Inference model to simulate emotional state inference and emotion concept learning, showing it can acquire emotion concepts and their learning depends on early experiences and attention patterns.

This paper develops a neuroethological model of stress and trauma using active inference to simulate adaptive and maladaptive behaviors in response to stressors.

This study evaluates the impact of global signal regression on DCM estimates of noise and effective connectivity in resting state fMRI, finding minimal effects on connectivity but caution in interpreting noise parameters.

This paper shows that biologically plausible cost functions in neural networks correspond to variational free energy under specific prior beliefs, enabling optimization of network parameters.

The paper shows that human hippocampal theta power increases during sequential spatial planning and correlates with decision speed and accuracy.

This paper compares two free energy functionals in active inference, showing how they differ in their treatment of future observations and policy selection.

This paper develops a deep active inference model to simulate neurocomputational mechanisms of emotional awareness and identifies potential clinical implications for personalized interventions.

This paper extends a computational model of hallucinations to explain in- and out-of-context hallucinations using active inference principles.

The paper presents two computational mechanisms for goal-directed exploration based on different types of information-gain, showing how they induce distinct patterns of Bayes-optimal behavior.