PLoS Computational Biology
November 7, 2007
Marco Loh, Edmund T Rolls, Gustavo Deco
172 citations
Reduced depth in the basins of attraction of cortical attractor states destabilizes neural activity at the network level due to constant statistical fluctuations from stochastic spiking of neurons. In integrate-and-fire network simulations, decreasing NMDA receptor conductances reduces attractor basin depth, destabilizes short-term memory states, and increases distractibility. Cognitive symptoms of schizophrenia—distractibility, working memory deficits, poor attention—could stem from this instability in prefrontal cortical networks. Lower firing rates in orbitofrontal and anterior cingulate cortex may account for negative symptoms like reduced emotions. Decreasing both GABA and NMDA conductances causes switches between attractor states and jumps from spontaneous activity into attractors, linked to positive symptoms such as delusions, paranoia, and hallucinations from shallow basins in temporal lobe semantic memory networks.
Frontiers in Psychology
January 1, 2020
Edmund T Rolls
43 citations
The global workspace hypothesis of consciousness faces problems, including how global the workspace must be for consciousness to emerge. Carruthers's (2019) version, which excludes conceptual representations and reduces phenomenal consciousness to physical processes, is also challenged. A different levels-of-explanation approach to the brain-mind relation is advocated. An alternative theory proposes a computational system using Higher Order Syntactic Thoughts (HOST) to perform credit assignment on first-order thoughts of multi-step plans by manipulating symbols in syntactic working memory. This provides an evolutionary reason for such a module, with which phenomenal consciousness is associated. The HOST approach offers advantages over global workspace and Higher Order Thought theories, and may involve prefrontal cortex regions, especially the left inferior frontal gyrus.
Frontiers in Computational Neuroscience
January 1, 2021
Edmund T Rolls
19 citations
Mental states and brain states are linked by a non-causal supervenient relationship, not by direct causation. Events at sub-neuronal, neuronal, and network levels occur simultaneously to perform computations describable as mental states with content about the world. Causality operates within levels of explanation but not between them, allowing mental properties to be emergent yet mechanistically expected. This theory avoids dualism and reductive physicalism, rooted in computational processes. For arithmetic, mental-level algorithmic descriptions are useful; for psychiatric disorders, understanding neural mechanisms aids treatment.
Frontiers in Computational Neuroscience
January 1, 2021
Edmund T Rolls
17 citations
A computational neuroscience theory of mind-brain relations proposes that mental states are high-level descriptions of simultaneous sub-neuronal, neuronal, and network-level computations. These levels are linked by non-causal supervenience, not causation; causality operates only within levels, not between them. The theory requires causality to satisfy three conditions: interventionist tests, same-level events, and a temporal order with a timescale of about 10 ms. While mental-level causal descriptions can be useful, brain-level accounts may be more accurate because mental-level accounts can involve confabulation. Cases of apparent downward causation are reinterpreted as within-level causation. This approach offers a path beyond Cartesian dualism and physical reductionism.
Brain Sciences
May 16, 2025
Edmund T Rolls
7 citations
A unified theory proposes that motivational states drive goal-directed actions to obtain anticipated rewards or avoid punishers, while emotional states arise when those rewards or punishers are actually received or not. The same genes and brain systems define primary rewards and punishers, like sweet taste or pain, and learn to anticipate them. The primate orbitofrontal cortex computes reward value, and the anterior cingulate cortex learns actions to obtain goals. Overlearned habits involve less emotion. The orbitofrontal cortex connects to language-related inferior frontal gyrus for declarative reports of subjective feelings. Reasoning brain systems offer alternative strategies for obtaining rewards or avoiding punishers, sometimes setting different goals than emotional systems.
Philosophia Scientiæ
January 19, 2026
Edmund T Rolls
3 citations
Computational neuroscience offers a new approach to classic philosophical problems—the mind-body problem, determinism and free will, and phenomenal consciousness. The mind and brain are different levels of explanation for information processing, and their correspondence can be understood through neural computation. However, a gap remains in understanding how brain events produce subjective experience. The proposed explanation is that 'feeling like something' arises from a computational process involving higher-order thoughts grounded in the world.
Advances in Consciousness Research
October 28, 2010
Edmund T Rolls
Noise from random neuronal spiking gives the brain advantages like probabilistic decision-making but makes it a non-deterministic system, raising implications for free will. Decisions are taken probabilistically between the reasoning system and the implicit reward system. The reasoning system's operation can be described as free will, and consciousness arises from using higher order syntactic thoughts (HOSTs) to correct first order thoughts. When the implicit system decides, we may confabulate a reason, making the feeling of free will an illusion.
Neural networks : the official journal of the International Neural Network Society
November 1, 2007
Edmund T Rolls
Recordings from populations of neurons in the inferior temporal visual cortex indicate that most information about which stimulus is shown comes from the firing rate of each neuron, not from stimulus-dependent synchrony, making it unlikely that synchrony or oscillations are essential for visual object perception. Backward masking experiments show the threshold for conscious visual perception is set higher than the level at which small but significant information exists in neuronal firing, allowing guessing without conscious awareness; this may prevent noise in sensory pathways from interrupting conscious thought processes.