Analog Cognition and Consciousness

A simplifying introduction

The brain’s currency has long been described as spikes zipping through synaptic circuits. But the cortex also generates something slower and larger: rhythmic electric fields — brain waves — that wash across millimeters of tissue. The analog cognition framework, proposed by Earl K. Miller, Scott L. Brincat and Jefferson E. Roy (MIT Picower Institute; Journal of Neuroscience, 2026), argues that these waves are not just timing signals or byproducts. In their account, synapses store what the brain knows, while wave dynamics decide — moment to moment — which stored patterns are expressed, and where waves overlap they physically add, subtract and reshape activity: an analog computation performed by the medium itself. Cognition is the flexible result; consciousness, the authors propose, is what happens when these wave patterns bring the cortex into one organized, globally integrated state.

The hybrid synapse/wave theory

• Core proposal: cognition and consciousness “may arise from bidirectional interactions between neuron spiking and large-scale brain waves” — with waves providing “an executive-like control process that organizes millions of neurons into lower-dimensional, coordinated patterns in real time”. ¹
• A feedback loop: synapses shape spiking; spiking generates electric fields (via ephaptic coupling); the fields, in turn, shape spiking — waves are both product and controller of neural activity.
• Why a second system is needed: individual neurons are multifunctional — they participate in many overlapping networks (“mixed selectivity”). Something must rapidly select which ensemble a given neuron belongs to at each moment; the framework argues the oscillatory fields do exactly that, letting the same cells take on “different roles depending on context” — a flexibility the authors call the “sine qua non of cognition”.

Brain waves organize neural information

• Waves both reflect and actively shape excitability (measurable as LFPs or EEG), with established roles in organizing information, routing it between circuits, and coordinating memory and control processes (e.g., segregating spiking to different oscillation phases).
• Frequency division of labor: alpha/beta rhythms (~13–30 Hz) carry top-down context, goals and memories and act as a “brake” (beta suppression precedes movement); gamma (~30–80 Hz) plus spiking carry feedforward sensory/motor “contents”. Slow rhythms modulate fast ones.
• Miller’s lab found this pattern forms a specific laminar organization — gamma strongest in superficial (feedforward) layers, alpha/beta deeper — conserved across four species (mouse, marmoset, macaque, human) and 14+ cortical areas.

Spatial computing

• The theory’s control mechanism: top-down alpha/beta activity is spatially structured — effectively a “mobile stencil” that governs where gamma/spiking may express information, by locally raising or lowering excitability. Sensory contents arrive via spatially diffuse feedforward/recurrent connectivity; the intersection of content and stencil determines which neurons participate.
• Consequently, the same stimulus can drive different spatial population patterns in different contexts — explaining mixed selectivity and dynamic, low-dimensional “subspace” coding as consequences of control in space and time (“spatiotemporal computing”), since waves also travel.

Traveling waves and representational change

• Brain waves typically sweep across the cortical surface (delta–gamma, ~1–80 Hz) as traveling waves, modulating the excitability of local neurons as they pass. Roles are reported in perception, attention and memory; propagation speeds are consistent with long-range horizontal connections in superficial cortex.
• Representational change: as a wave travels, which neurons are momentarily excitable shifts with it — so wave position is linked to which representation (e.g., which item in a remembered sequence) is currently expressed.

Analog computation with waves

• Where waves intersect they superpose — adding, cancelling, filtering. Variables can be encoded as amplitude and phase of the waves, whose interactions compute “everywhere at once”: inherently parallel, rather than the sequential, step-by-step style of digital computation (the review compares mechanical tide predictors and differential analyzers).
• Efficiency: analog field interactions avoid metabolically expensive all-or-none spiking at every step; the review notes neuromorphic analog hardware can gain more than an order of magnitude in energy efficiency over digital equivalents.
• Open question: direct signatures of analog computation in the brain have not yet been demonstrated — “We aim to test it by looking for signatures of analog computation in brain wave patterns” (Miller, via MIT News). ²

Consciousness

• The framework’s definition: “Consciousness, then, emerges when these dynamic wave patterns bring the cortex into an organized, globally integrated state, one that naturally links and influences widespread activity.”
• Anesthesia evidence: propofol (GABAergic), ketamine (NMDAergic) and dexmedetomidine (α2-adrenergic) act on different molecular targets yet converge on the same systems-level outcome — high-power, slow (1–4 Hz delta), temporally misaligned waves that disrupt coordinated communication. Anesthetics do not simply shut the cortex off. Conclusion: “consciousness depends less on specific receptors or cell types and more on the integrity of large-scale wave organization.”
• Relation to other theories: it shares with Global Workspace / GNWT a commitment to large-scale integration — but locates the mechanism in dynamic field interactions rather than information broadcasting; shares with IIT an emphasis on organization and self-influence — but places it in large-scale wave dynamics rather than a “hot zone”; and it aligns with Higher-Order Theories and predictive coding, reframing consciousness as arising when top-down, wave-implemented models impose organized structure on lower-level cortical processes (cf. the “nested observer” model). See active-inference for the predictive-processing framework this connects to.

Clinical outlook and limitations

• Because electric fields are measurable and can be manipulated non-invasively, the framework suggests therapeutic routes — “Developing treatments based on brain wave dynamics is not just an opportunity, but also an obligation” (Miller); the lab is part of a collaboration studying brain waves in autism.
• Limitations: this is a theory/review synthesis (“may arise”), not yet direct proof. The key prediction — signatures of analog computation in wave patterns — remains untested; and the functional weight of field effects in tissue is subtle and still actively debated (see ephaptic-coupling-and-brain-waves).

Relationship to This Wiki

• ephaptic-coupling-and-brain-waves — companion page: the physical mechanism by which waves directly influence spiking.
• spiking-neural-networks — the spike-based view of neural computation that this framework complements with an analog, field-based substrate.
• active-inference — predictive-processing/active inference; the analog theory explicitly aligns with predictive coding of cortical function.

References

• Miller, E. K., Brincat, S. L., & Roy, J. E. (2026). “Analog Cognition and Consciousness.” Journal of Neuroscience, 46(33), e0711262026. DOI · journal page. Author manuscript (full text) retrieved from OSF. ¹
• Orenstein, D. (2026). “Cognition and consciousness arise from analog computations, says new theory.” MIT News, 1 September 2026. Link. ²
• Key primary literature cited in the review (full list in the raw capture): Katz & Schmitt (1940) J. Physiol.; Anastassiou et al. (2011) Nat. Neurosci.; Fröhlich & McCormick (2010) Neuron; Lundqvist et al. — Spatial Computing model; Bardon et al. (2025) and Eisen et al. (2026) — anesthetic convergence on wave changes.

Footnotes

1. raw/papers/miller-2026-analog-cognition-consciousness.md ↩ ↩²

2. raw/articles/mit-news-analog-cognition-2026.md ↩ ↩²