Resourceful Event-Predictive Inference: The Nature of Cognitive Effort

Frontiers in Psychology, Journal Year: 2022, Volume and Issue: 13

Published: June 30, 2022

Pursuing a precise, focused train of thought requires cognitive effort. Even more effort is necessary when alternatives need to be considered or the imagined situation becomes complex. Cognitive resources available us limit we can spend. In line with previous work, an information-theoretic, Bayesian brain approach pursued: solve tasks in our environment, needs invest information, that is, negative entropy, impose structure, focus, away from uniform structure other task-incompatible, latent structures. To get complete formalization effort, resourceful event-predictive inference model (REPI) introduced, which offers computational and algorithmic explanations about generative models, active dynamics unfold within, required steer dynamics—to, for example, purposefully process sensory signals, decide on responses, invoke their execution. REPI suggests infer preparatory priors, activate anticipate action consequences. Due limited resources, though, are prone task-irrelevant distractions. For side imperative stimulus causes Simon effect and, due similar reasons, fail optimally switch between tasks. An actual implementation simulates such task interactions first estimates involved The may further studied promises offer deeper why quickly exhausted multitasking, how influenced by irrelevant modalities, exhibit magnitude interference, during social interactions, often take perspective others into account.

Language: Английский

---

Model-Based Approaches to Investigating Mismatch Responses in Schizophrenia

Clinical EEG and Neuroscience, Journal Year: 2024, Volume and Issue: unknown

Published: May 15, 2024

Alterations of mismatch responses (ie, neural activity evoked by unexpected stimuli) are often considered a potential biomarker schizophrenia. Going beyond establishing the type observed alterations found in diagnosed patients and related cohorts, computational methods can yield valuable insights into underlying disruptions mechanisms cognitive function. Here, we adopt typology model-based approaches from neuroscience, providing an overview study their schizophrenia four complementary perspectives: (a) connectivity models, (b) decoding (c) network (d) models. Connectivity models aim at inferring effective patterns between brain regions that may underlie measured sensor level. Decoding use multivariate spatiotemporal response to infer sensory violations or classify participants based on diagnosis. Neural such as deep convolutional networks be used for improved classification performance well systematic various aspects empirical data. Finally, quantify terms signaling updating perceptual predictions over time. In addition describing available methodology reviewing results recent psychiatry studies, offer suggestions future work applying techniques advance

Language: Английский

---

High-level visual prediction errors in early visual cortex

PLoS Biology, Journal Year: 2024, Volume and Issue: 22(11), P. e3002829 - e3002829

Published: Nov. 11, 2024

Perception is shaped by both incoming sensory input and expectations derived from our prior knowledge. Numerous studies have shown stronger neural activity for surprising inputs, suggestive of predictive processing. However, it largely unclear what predictions are made across the cortical hierarchy, therefore kind surprise drives this up-regulation activity. Here, we leveraged fMRI in human volunteers deep network (DNN) models to arbitrate between 2 hypotheses: prediction errors may signal a local mismatch expectation at each level or be computed higher levels resulting broadcast earlier areas hierarchy. Our results align with latter hypothesis. Prediction low- high-level visual cortex responded high-level, but not low-level, surprise. This scaling early strongly diverged feedforward tuning. Combined, suggest that constrain processing areas, thereby aiding perceptual inference.

Language: Английский

---

Using multi-modal neuroimaging to characterise social brain specialisation in infants

eLife, Journal Year: 2023, Volume and Issue: 12

Published: Oct. 11, 2023

The specialised regional functionality of the mature human cortex partly emerges through experience-dependent specialisation during early development. Our existing understanding functional in infant brain is based on evidence from unitary imaging modalities and has thus focused isolated estimates spatial or temporal selectivity neural haemodynamic activation, giving an incomplete picture. We speculate that will be underpinned by better coordinated metabolic changes a broadly orchestrated physiological response. To enable researchers to track this process development, we develop new tools allow simultaneous measurement activity (EEG), rate, oxygenated blood supply (broadband near-infrared spectroscopy) awake infant. In 4- 7-month-old infants, use these show social processing accompanied spatially temporally specific increases coupled activation temporal-parietal junction, core hub region adult brain. During non-social processing, decreased same region, indicating specificity processing. Coupling was strongest with high-frequency (beta gamma), consistent greater energetic requirements more localised action activity. development multimodal measures future open vistas

Language: Английский

---

The Controllosphere: The Neural Origin of Cognitive Effort

Published: March 26, 2023

Why do some mental activities feel harder than others? The answer to this question is surprisingly controversial. Current theories propose that cognitive effort affords a computational benefit, such as instigating switch from an activity with low reward value different higher value. By contrast, in article I relate the fact brain neuroanatomy and neurophysiology render neural states more energy-efficient others. introduce concept of “controllosphere,” energy-inefficient region state space associated high control, which surrounds better-known “intrinsic manifold”, subspace control. Integration control-theoretic principles classic neurocomputational models control suggests dorsolateral prefrontal cortex (DLPFC) implements controller can drive system into controllosphere, anterior cingulate (ACC) observer monitors changes controlled system, reflects mismatch between DLPFC ACC energies for observation. On account, scales energetic demands signal, especially when consequences are unobservable by ACC. Further, transitions through controllosphere lead buildup waste. Cognitive therefore prevents against damage discouraging extended periods

Language: Английский

---