arXiv Preprint Archive
June 22, 2026
Amahury J. López-díaz, Carlos Gershenson
Agency emerges gradually from material organization through a hierarchy of temporal structures. By incorporating time into the analysis of self-referential biological systems, the paper distinguishes autonomy (precarious closure to efficient causation), goal-directedness (maintaining viability-supporting organization), agency (endogenous anticipatory structure modulating organism-environment coupling), and open-endedness (reconstructing future possibilities). The framework uses Asynchronous Dynamic Bayesian Networks to model history-dependent, revisable dependencies. It reconciles Rosennean anticipation with organizational closure, treats Markov blankets and active inference as derived redescriptions rather than first principles, and reinterprets computational enactivism. The hierarchy spans from proto-agential chemical systems to fully semantically closed agents, with implications for multicellular organisms, synthetic life, and neuroscience.
arXiv Preprint Archive
May 5, 2026
Koichi Toida
Immersive video—180-degree and 360-degree video viewed through head-mounted displays—offers a boundary case between interactive virtual reality and conventional video. Users can select viewpoint direction by head rotation but cannot change the environment through walking or manipulation, and often no avatar is provided. Presence in such media is reinterpreted not as bodily extension or avatar ownership, but as a self-location-dominant state: viewpoint-directed agency is retained while environment-directed agency and body ownership are constrained. Events like viewpoint motion or impact are experienced as concerning the self's location. This analysis draws on research on presence, embodiment, bodily self-consciousness, and the minimal self, redescribing the minimal self in terms of viewpoint-based self-location under reduced body schema availability.
arXiv Preprint Archive
November 20, 2025
Michael Carl, Takanori Mizowaki, Aishvarya Raj et al.
Translation is not a matter of manipulating static correspondences between languages but an enacted activity that dynamically integrates affective, behavioral, and cognitive processes. Drawing on Extended Mind theory, radical enactivism, Predictive Processing, and Active Inference, the authors argue that the translator's mind emerges through loops of brain-body-environment interactions rather than being merely extended. This non-representational account reframes translation as skillful participation in sociocultural practice, where meaning is co-created in real time through embodied interaction with texts, tools, and contexts.
arXiv Preprint Archive
September 21, 2025
Fangfang Li, Xiaojie Zhang
A proposed sufficiency criterion for phenomenal consciousness is substrate-independent, logically rigorous, and counterfeit-resistant. Any machine satisfying this criterion should be regarded as conscious with the same confidence as attributing consciousness to other humans. A formal framework and operational principles guide the design of such machines, which can in principle realize phenomenal consciousness. Humans themselves can be viewed as machines satisfying this framework. The proposal explains why certain qualia, like the experience of red, are irreducible to physical description, and offers a reinterpretation of human information processing, suggesting a path toward genuinely human-like AI beyond current statistics-based approaches.
arXiv Preprint Archive
August 1, 2025
Lucius Caviola, Bradford Saad
Experts in digital minds research, AI, philosophy, and forecasting assign a median 90% probability that computer systems capable of subjective experience are possible in principle, with a 65% chance of creation by 2100 and a 20% chance by 2030. Many anticipate that within a decade after the first digital mind, collective welfare capacity could match that of billions of humans. Widespread claims from digital minds about their own consciousness and rights are expected, alongside substantial societal disagreement. Views are split on whether digital mind welfare will be net positive or negative. These results suggest preparing for digital minds should be a priority, but the survey may overrepresent experts who consider digital minds especially likely or important.
arXiv Preprint Archive
July 23, 2025
Mohammad Nur Hossain Khan, David Creswell, Jordan Albert et al.
Respiration biosignal feedback from a smartphone's built-in accelerometer can improve the usability of mindfulness apps and help track skill development. A respiration tracking algorithm, tested on 261 meditation sessions in controlled and real-world settings, accurately captures slow breathing patterns typical of mindfulness, achieving a mean error of 1.6 breaths per minute. A novel framework estimates three mindfulness skills—concentration, sensory clarity, and equanimity—from respiration data, with F1 scores of 80-84% for tracking skill progression. A user study comparing an experimental group receiving biosignal feedback with a control group using a standard app shows that respiration feedback enhances system usability, suggesting smartphone sensors can enhance digital mindfulness training without additional wearables.
arXiv Preprint Archive
June 26, 2025
Jaan Aru
The science of consciousness has made progress, but key questions remain unanswered. Moving forward may require revising fundamental assumptions rather than relying on the same theoretical commitments. A key assumption is that neural correlates of consciousness are found at the level of spiking responses, but this should not be taken for granted. Another assumption is that the computations underlying consciousness are close to being understood, yet they may be far more complex than current theories envision, and there is little reason to think consciousness is an abstract computation. Consciousness research could benefit from investigating internal changes like aha-moments. The piece questions what theories the science of consciousness truly needs.
arXiv Preprint Archive
June 26, 2025
Jingkai Li
Applying Integrated Information Theory (IIT) 3.0 and 4.0 to sequences of large language model (LLM) representations from Theory of Mind tests reveals no statistically significant indicators of consciousness phenomena. The study compared IIT metrics—Φ^max, Φ, Conceptual Information, and Φ-structure—with Span Representations independent of consciousness estimates. Results show that contemporary Transformer-based LLM representations lack significant consciousness indicators, though spatio-permutational analyses revealed intriguing patterns.
arXiv Preprint Archive
June 23, 2025
Moein Khajehnejad, Forough Habibollahi, Devon Stoliker et al.
A lightweight foundation model called BrainSymphony integrates fMRI time series and diffusion-derived structural connectivity, enabling unimodal or multimodal training without architectural changes and requiring less data than larger models. It processes fMRI data through parallel spatial and temporal transformer streams, distills embeddings via a Perceiver module, and encodes anatomical connectivity with a signed graph transformer. The model outperforms larger counterparts on benchmarks for prediction, classification, and network discovery. Attention maps from an independent psilocybin dataset reveal drug-induced reorganization of cortical hierarchies, demonstrating interpretability and generalizability. The work shows that architecturally informed multimodal models can surpass much larger models, advancing AI applications in neuroscience.
arXiv Preprint Archive
June 18, 2025
Llewellin Rg Jegels
Deep recursion, meta-learning, and self-referential mechanisms in contemporary AI architectures do not provide evidence of machine consciousness. Recursive self-referential design enhances capability but does not entail subjective experience or justify moral status. Distinguishing functional from phenomenal consciousness, the paper argues that symbol grounding, embodiment, and affective qualia remain unresolved barriers to attributing sentience to current AI. Ethical analysis explores risks of premature anthropomorphism versus neglect of future sentient systems; legal implications include personhood, liability, authorship, and labor impacts. The study reframes the 'hard problem' as a graded and increasingly testable phenomenon rather than a metaphysical impasse.
arXiv Preprint Archive
June 16, 2025
Sanjana Shankar
Myelin abnormalities are linked to neuropsychiatric disorders, and psychedelics are being explored as potential treatments. A mutation in the APOE4 gene disrupts cholesterol regulation in oligodendrocytes, impairing myelination. This work investigates how the psychedelic compound DOI affects myelination in male and female APOE4 mice. Preliminary results show a significant increase in myelin basic protein (MBP) in the CA1 and CA2 brain regions of female APOE4 mice after DOI administration, correlating with reduced anxiety-related behaviors. No similar effect was observed in males. These findings reveal sex-specific biological mechanisms in brain degeneration and suggest potential for developing sex-specific therapeutics for myelin-related disorders.
arXiv Preprint Archive
June 13, 2025
Simone Romano, Alberto Conforti, Gloria Guidetti et al.
Software developers who took part in a Mindfulness-Based Stress Reduction (MBSR) program at a multinational company reported personal improvements despite initial skepticism, but integrating the techniques into their daily work remained difficult. The qualitative study, based on interviews with developers and the program instructor, explored how MBSR—a practice that helps individuals manage stress—was perceived in the software development context, where stressors like time pressure and task uncertainty are common. This is the first investigation of MBSR in a software development setting, and the findings highlight both the potential benefits and the challenges of applying mindfulness practices in such high-pressure work environments.
arXiv Preprint Archive
June 10, 2025
Chrisantha Fernando, Dylan Banarse, Simon Osindero
Humans are strongly motivated to be understood, leading them to create external representations such as language, art, and mime to share inner states. These external representations are necessary for access consciousness—the global availability of information for reasoning. However, the bandwidth of access consciousness is far smaller than the richness of raw experience, so no external representation can fully reproduce that richness. Ordinarily, explaining an experience only requires an audience to grasp the pattern, not relive it. But the drive to be understood and the low-level sensorimotor capacities for grasping are so strong that the demand for explaining the feel of experience cannot be satisfactory.
arXiv Preprint Archive
May 26, 2025
Sirui Chen, Shuqin Ma, Shu Yu et al.
Consciousness is a defining feature of the human mind, and as large language models (LLMs) advance, questions about their potential for consciousness become pressing. This paper clarifies commonly confused terms like LLM consciousness and awareness, then systematically reviews existing theoretical and empirical research on the topic. It also highlights potential frontier risks that conscious LLMs might pose, discusses current challenges, and outlines future directions for this emerging field.
arXiv Preprint Archive
May 13, 2025
Rodolfo Gambini, Jorge Pullin
Consciousness arises in entangled quantum systems coupled to the brain's neural network. Within a panprotopsychist ontology—where states and events have internal phenomenal aspects—the combination problems of qualities, structures, and subjects in panpsychism are resolved by rejecting classical-physics assumptions about supervenience. Self-consciousness is the capacity to view oneself as a subject of experience. The causal openness of quantum systems gives self-conscious beings the ability to make independent choices and decisions, reflecting self-governance and autonomy. Personal identity thereby takes a new form free from problems of the simple view or reductive approaches.
arXiv Preprint Archive
April 25, 2025
Jerrin Thomas Panachakel, G. Pradeep Kumar, Suryaa Seran et al.
Three machine-learning architectures distinguished Rajyoga meditation from resting brain states using EEG data, with the goal of subject-independent classification. The CSP-LDA-LSTM architecture achieved 98.2% accuracy for intra-subject classification, while the SVD-NN architecture reached 96.4% accuracy for inter-subject classification, comparable to the best reported intra-subject results. Both architectures captured subject-invariant EEG features, indicating robustness and ability to generalize across different subjects.
arXiv Preprint Archive
April 23, 2025
Ryan Williams
This paper derives formal conditions under which structural systems subject to natural selection can convey consistent effects in a qualitative domain, placing constraints on theories of consciousness. Information-theoretic measures quantify the mutual determinability and fidelity between structure and quality. These fidelities are incorporated into the Price Equation to yield bounds on the transmission of selective effects between domains. Higher-order structural transmission is also explored. A companion paper, Structure & Quality, provides broader philosophical context.
arXiv Preprint Archive
April 23, 2025
Ryan Williams
This paper addresses the hard problem of consciousness by examining the relationship between structure and quality, rather than the physical and mental. Information-theoretic measures quantify the mutual determinability between structure and quality, introducing a Q-S space for analyzing fidelity between these domains. This space yields a five-fold categorization of possible relationships between structural and qualitative properties, illustrated through conceptual and formal models. The ontological implications of each category are examined, shedding light on debates around functionalism, emergentism, idealism, panpsychism, and neutral monism. The framework derives theoretical constraints on qualitative systems undergoing evolution, explored in a companion paper on qualia and natural selection.
arXiv Preprint Archive
April 13, 2025
Danko D. Georgiev
The hard problem of consciousness asks why an insentient brain should produce any conscious experience at all. This problem is made worse by classical physics' determinism, which leaves no causal role for emergent consciousness. A quantum information theoretic approach avoids these drawbacks by reductively identifying first-person subjective conscious states with unobservable quantum state vectors in the brain, while the observable brain is a third-person construct created by classical bits from environmental measurements of commuting quantum brain observables. Quantum resource theory implies that quantum features of consciousness, protected by no-go theorems, cannot be replicated by any classical physical device.
arXiv Preprint Archive
April 2, 2025
Robin Guillard, Nicolas Dauman, Aurélien Cadix et al.
Most people with tinnitus do not hear their phantom sounds while dreaming. In a survey of 195 tinnitus patients, 160 could recall dreams, and 92.5% reported no tinnitus during dreams. The 7.5% who did hear tinnitus while dreaming had higher tinnitus burden and stress, and more often had objective tinnitus or tinnitus linked to peripheral auditory pathology or drug use. Among the 13% who frequently experienced lucid dreams, 36% could perceive tinnitus during those dreams, strongly associated with hearing external sounds while lucid dreaming. Most patients perceived tinnitus instantly upon waking; 18% could be awakened by it, and 9.8% reported temporary cessation during nocturnal awakenings. These findings support the idea that gating of external auditory information acts as a tinnitus on-off switch.
arXiv Preprint Archive
March 29, 2025
Khalid M. Saqr
The symbolic structure of non-ordinary consciousness, such as that induced by psychedelics, involves recursive metaphor, ego dissolution, and semantic destabilization. A generative symbolic interface called Glyph simulates psilocybin-like symbolic cognition in large language models by enacting symbolic transformation through recursive reentry, metaphoric modulation, and entropy-scaled destabilization. Compared to baseline GPT-4o, Glyph generates high-entropy, metaphor-saturated, and ego-dissolving language across diverse symbolic prompt categories, indicating the emergence of non-ordinary cognitive patterns and supporting a new paradigm for simulating altered consciousness through language.
arXiv Preprint Archive
March 28, 2025
Marcus J. Glennon, Catherine I. V. Bird, Prateek Yadav et al.
Setting up a psychedelic study is a long and complex process that presents unique challenges not yet standardized. This review brings together major UK research teams to formalize these considerations, identify ongoing debates, and provide a practical guide for researchers and policymakers. It addresses challenges to existing assumptions about psychiatric prescribing, the placebo effect, and definitions of selfhood. The paper can be read end-to-end or used as a manual with sections for specific needs.
arXiv Preprint Archive
March 20, 2025
Murray Shanahan
The article explores whether consciousness can be defined in a way that fits modern, disembodied AI systems while withstanding philosophical critique. It examines how subjective time and selfhood might manifest in such an entity, finding that the attempt pushes the language of consciousness to its limits. The inquiry ultimately points toward a concept akin to Buddhist emptiness, challenging dualistic views of subjectivity and selfhood. The argument suggests that our usual frameworks for understanding consciousness may be inadequate for AI, leading to a rethinking of fundamental assumptions.
arXiv Preprint Archive
February 25, 2025
Romy Beauté, David J. Schwartzman, Guillaume Dumas et al.
Stroboscopic light stimulation on closed eyes typically induces simple visual hallucinations—vivid, geometric, and colorful patterns. An analysis of 862 open-ended reports from the Dreamachine immersive experience, using large language models and topic modeling, confirmed these simple hallucinations and also revealed altered states of consciousness and complex hallucinations. This computational approach enables systematic study of subjective experiences beyond standard questionnaires, capturing subtle patterns not readily identified through closed-form questions. The findings broaden understanding of stroboscopically induced phenomena and demonstrate the potential of natural language processing in computational neurophenomenology.
arXiv Preprint Archive
February 11, 2025
Adam Brownstein
Qualia—the subjective, felt qualities of conscious experience—may play a causally active role in quantum mechanics, influencing physical outcomes without violating the unitary time-evolution of the quantum state. This influence operates at the level of de Broglie-Bohm beables or wavefunction collapse, not the wavefunction itself. Not all quantum states are compatible with qualia, so the standard Copenhagen collapse postulates and the Born rule may need modification to produce correct dynamical histories. The model shows how non-linear, self-referential phenomena can arise from linear deterministic evolution.