A new study led by IFISC (CSIC-UIB, Unidad de Excelencia María de Maeztu) has uncovered how the brain flexibly switches communication pathways by modulating the balance between two fundamental inhibitory mechanisms. Researchers combined computational modeling with experimental hippocampal recordings to show how interactions between slow theta oscillations and fast gamma rhythms can be directed either from fast to slow or vice versa, depending on the dominance of feedforward or feedback inhibition. This mechanism, presented in PLoS Computational Biology, provides a dynamic way for neural circuits to prioritize and route information, with direct implications for memory, learning, and attention.
Traditionally, neuroscientists believed that slow brain rhythms, such as theta, organize faster activity like gamma oscillations. However, this new study demonstrates that this relationship is bidirectional. Using a theoretical framework that integrates electrophysiological data from rats exploring novel and familiar environments, the researchers identified two operational modes: in one, feedforward inhibition leads to gamma-to-theta interactions, while in the other, feedback inhibition produces theta-to-gamma interactions. Importantly, real neural circuits combine both motifs, and the transition between them can be smoothly tuned by synaptic strength within biologically realistic ranges.
“This work provides a mechanistic explanation for how the brain flexibly shifts communication channels depending on context”, states Claudio Mirasso, IFISC researcher and author of the study. “By adjusting the balance between different types of inhibition, circuits can decide which inputs to prioritize, whether they come from memory-related pathways or from new sensory information”.
The team validated their model using pathway-specific recordings from the hippocampus. They found that when animals explored familiar environments, circuits favored a feedback-dominated mode, enhancing direct transmission from entorhinal cortex to hippocampus. By contrast, novelty exploration induced a shift toward feedforward-dominated interactions, allowing parallel pathways to contribute more strongly and integrating memory retrieval with incoming sensory inputs. This flexible switching mechanism suggests that cross-frequency interactions are not fixed but adapt to behavioral demands.
“Our results help unify competing views on how cross-frequency coupling emerges”, explains Mirasso. “Instead of being purely local or inherited from upstream regions, these rhythms emerge from an interplay between external inputs and local inhibitory dynamics. This dual mechanism allows the brain to optimize information processing under different conditions”.
Beyond memory and navigation, the findings may extend to other cognitive functions. A recent human electrocorticography study on attention revealed patterns consistent with the model, indicating that the same principles may govern how the brain filters relevant from distracting stimuli. By providing a conceptual framework that links inhibition balance to oscillatory dynamics, the study opens new avenues for investigating brain flexibility across regions and tasks.
Looking ahead, the researchers aim to expand their minimal model to include a richer diversity of interneuron types and region-specific architectures. Such refinements could shed light on clinical conditions where cross-frequency coupling is disrupted, such as epilepsy, schizophrenia, or Alzheimer’s disease. “Understanding these dynamics at a mechanistic level may eventually inspire new strategies for therapeutic intervention”, concludes Mirasso.
Photo: From left to right, Santiago Canals (Instituto de Neurociencias de Alicante), Dimitrios Chalkiadakis (IFISC, UIB-CSIC) and Claudio Mirasso (IFISC, UIB-CSIC). Comuniación Instituto de Neurociencias CSIC-UMH
Reference
Chalkiadakis D, Sánchez-Claros J, López-Madrona VJ, Canals S, Mirasso CR (2025). The role of feedforward and feedback inhibition in modulating theta-gamma cross-frequency interactions in neural circuits. PLoS Comput Biol 21(8): e1013363. https://doi.org/10.1371/journal.pcbi.1013363