Traveling Brain Waves: Collective Computation and Electrical Feedback

Brain waves aren't just an echo of neurons firing — they might be doing computational work themselves, feeding back to shape the very activity that produced them.

The usual picture is that neurons compute and traveling waves are a byproduct, a kind of visible ripple on top of the "real" processing. Ruffini argues for something different: the wave itself is a collective mode of the computation, a population-level pattern that helps coordinate what individual neurons do next. He extends the analogy of electronic oscillators that synchronize by computing an average frequency and then using that shared rhythm to steer themselves — except instead of one global clock, a traveling wave is a moving coordination pattern sweeping through the network. In behaving primates, the timing and position of these waves actually predicts whether a stimulus gets detected, which is the empirical anchor for treating waves as functionally load-bearing rather than epiphenomenal.

The mechanism for this feedback is electric fields generated by population activity, acting back on neurons without going through a synapse — so-called ephaptic coupling. The natural objection is that these endogenous fields are far too weak (fractions of a volt per meter) to do anything to neurons that aren't already about to fire. Ruffini pushes back using slice and awake-macaque data: fields in the 0.2–0.3 V/m range, comparable to what transcranial electrical stimulation delivers, measurably shift spike timing and can even block or redirect wave propagation. That doesn't prove endogenous fields matter much for everyday cognition — it just kills the blanket claim that they're too weak to matter at all. Notably, his earlier modeling work found that these naturally occurring mesoscopic fields resemble applied stimulation fields in strength and spatial character, which is why studying artificial brain stimulation can inform theories about the brain's own internal signaling.

The more speculative move ties this to Kolmogorov theory, Ruffini's framework linking structured conscious experience to models that compress information about the world and self. His hypothesis is that these fields belong inside the relevant computational system whenever they causally shape its models — not that fields are consciousness, just that they could be legitimate parts of the machinery. The proposed test is refreshingly falsifiable: run calibrated simulations with real field feedback, with feedback removed, and with feedback scrambled but equal in strength, then check whether wave organization, information integration, and task performance actually differ. A positive result would show a computational role within the model, though the source is explicit that this alone wouldn't establish anything about experience itself.