Three scientists at MIT’s Picower Institute for Learning and Memory propose a new theory, published in the Journal of Neuroscience, that explains how the brain generates cognition and consciousness: traveling rhythmic neural waves coordinate flexible networks to perform analog computations.
Picower Professor Earl K. Miller argues that the common “circuit” metaphor is incomplete. While physically connected circuits and synapses store memories and represent goals, they cannot alone support the rapid, improvisational processing required in constantly changing sensory contexts. Instead, the brain needs a control system that can synchronize millions of neurons within fractions of a second. Decades of experimental evidence show that brain waves—synchronized rhythmic fluctuations of large neuronal populations—fulfill this role.
Unlike digital circuits that compute step‑by‑step through sequential gates, analog computation exploits wave interference to process many operations in parallel, offering higher efficiency. Locally traveling waves are ubiquitous in the brain, reflecting the brain’s exploitation of its own physics for computation.
Clinically, the theory is attractive because waves can be modulated non‑invasively, opening new therapeutic avenues. Miller’s lab, in collaboration on autism research, is already exploring wave‑based interventions.
To argue that waves organize neurons, the authors start from the well‑established “mixed selectivity” of neurons: a single cell participates in multiple functional networks. Different frequency bands govern distinct cognitive processes: slower α/β waves encode memories and goals, while faster γ waves represent incoming sensory data. α/β control waves arise from coordinated spiking in synaptic circuits, yet the waves themselves rapidly influence spiking via ephaptic (electric‑field) coupling, creating a bidirectional loop.
α/β waves also act spatially, affecting local cortical regions, and temporally, traveling across the cortex. They function as mobile stencils that dictate where and when γ waves can process information and which neuronal ensembles are recruited. The authors term this “spatiotemporal computing”; where waves intersect they add and subtract, enabling analog operations.
Regarding consciousness, the paper posits that it emerges when dynamic wave patterns bring the cortex into a globally integrated state. Anesthesia studies by Miller and colleague Emery N. Brown show that three pharmacologically distinct agents all disrupt wave dynamics to induce unconsciousness, suggesting that consciousness depends more on large‑scale wave organization than on specific receptors or cell types.
The work was funded by the Freedom Together Foundation, the Picower Institute, the U.S. Army Research Office, the Office of Naval Research, a MURI grant, NIH, and the Simons Center for the Social Brain.
Blogger's Review: This perspective reframes the brain as an analog computing system that leverages intrinsic wave physics. Direct experimental validation of analog signatures would not only deepen our understanding of cognition but could also bridge neuroscience with emerging analog hardware technologies.