In a groundbreaking new study, researchers from the International Brain Laboratory have unveiled a comprehensive brain-wide map of neural activity, providing unprecedented insights into how choice is represented and formed across multiple regions during complex behavior. Published in Nature, this work transcends previous localized studies by revealing a distributed network of cortical and subcortical areas that collectively participate in encoding decisions, emphasizing the integral role of subcortical structures long overshadowed by cortical investigations.
Decades of research have highlighted choice-related neural activity primarily within frontoparietal regions of the primate cortex. These cortical areas exhibit ramping neuronal firing patterns consistent with the accumulation of sensory evidence leading to a behavioral choice. However, the present study expands this paradigm by focusing on rodents, leveraging advanced population decoding techniques and single-cell analyses to track the emergence and dynamics of choice signals in an array of brain regions, both cortical and subcortical.
By decoding neural population activity within a narrow 100-millisecond window just prior to movement onset, the researchers identified widespread representations of upcoming left versus right choices in numerous brain territories. Remarkably, some of the strongest choice signals were located deep within the brainstem’s hindbrain regions, including the gigantocellular reticular nucleus (GRN), pontine nuclei (PRNr), and medial accessory reticular nucleus (MARN). Alongside these areas, the thalamus, midbrain nuclei such as the substantia nigra pars reticulata (SNr), and the hypothalamus also exhibited pronounced choice-related modulations.
The study’s use of rigorous single-cell statistical models corrected for confounding factors such as stimulus presentation and task block effects, underscoring that a greater fraction of neurons responded specifically to choice direction than to the sensory stimuli themselves. This finding challenges traditional views that cortical sensory processing regions dominate the decision formation process, instead positioning multiple subcortical areas as critical hubs for choice coding.
Among these subcortical territories, the GRN stood out as a prime example of a region where individual neurons exhibited robust selectivity for right versus left choices. This choice preference was clearly delineated both in spike raster plots and in model-based encoding predictions, which faithfully captured variance in firing rates attributable solely to choice parameters. Population trajectory analyses further revealed that neural ensembles within the GRN evolved along distinct paths in state-space corresponding to each choice, highlighting a dynamic and gradual separation well before the animal’s first movement.
Furthermore, this population-level choice encoding was also strong in other hindbrain nuclei such as the intermediate reticular nucleus (IRN) and pontine reticular nucleus caudalis (PRNc), as well as in midbrain structures including the mesencephalic reticular nucleus (MRN) and superior colliculus (SCm). These results suggest that decision-related signals are not exclusively the province of higher cortical circuits but emerge from a concerted interaction of multiple, often evolutionarily conserved, brain regions.
Temporal analyses of choice signal latency demonstrated that some of the earliest neural differentiation between left and right choices appeared nearly simultaneously in both thalamic nuclei and cortical visual areas. This near synchrony implies a rapid, distributed initiation of decision signals rather than a strictly hierarchical, sequential processing model. Subsequently, a broader network encompassing numerous cortical and subcortical centers became engaged, with particular emphasis on brainstem reticular formation structures involved in motor preparation and execution.
Intriguingly, the study also found that movement onset represented distinct patterns of neural encoding across regions depending on the timing of behavioral responses. For example, in trials with early movements, certain subcortical structures like the periaqueductal gray (PAG) contributed significantly to model fits, whereas their influence waned in late-response trials. Conversely, secondary visual areas and motor cortical regions maintained consistent involvement regardless of response latency, suggesting differential recruitment dynamics across the brain depending on behavioral context.
The inclusion of cerebellar nuclei such as the central lobule (CENT2) in the ensemble of regions with strong choice encoding highlights the cerebellum’s emergent role in decision-making circuits, beyond its classical attribution to motor coordination. These findings align with growing evidence that cerebellar outputs participate in cognitive functions and behavioral planning.
Collectively, this extensive dataset paints a holistic picture of choice representation as a multifaceted phenomenon, wherein cortical sensory and motor regions interface with subcortical hubs to shape action plans. Such distributed coding likely affords the brain the flexibility and robustness needed for rapid, context-dependent decision-making in complex environments.
This study’s methodological sophistication—combining high-density electrophysiology, causal modeling, and advanced population decoding—sets a new standard for dissecting brain-wide neural dynamics during behavior. It directly confronts long-standing assumptions regarding the primacy of cortical circuits in decision formation and opens up promising avenues for exploring how subcortical regions contribute causally to behavior.
Understanding this distributed network’s precise mechanisms may have profound implications for neurological disorders where decision-making or motor control is compromised. By targeting subcortical nuclei implicated in choice formation, future therapeutic strategies could be more finely tuned to restore or modulate decision-related neural activity.
This landmark work exemplifies how collaborative, large-scale neuroscience efforts can unravel the complexity of neural computations underpinning behavior. As technological and analytical tools continue to advance, the prospect of fully mapping the brain’s decision circuitry with even greater resolution and causal specificity appears within reach.
In sum, the International Brain Laboratory’s study ushers in a new era of brain-wide investigation, revealing the intricate, multi-regional choreography of neurons that orchestrate choice, movement preparation, and execution. These findings underscore the importance of looking beyond traditional cortical territories, recognizing the concerted action of diverse brain areas that together enable adaptive, goal-directed behavior.
Subject of Research: Neural representation of choice across distributed brain regions during behavior
Article Title: A brain-wide map of neural activity during complex behaviour
Article References:
International Brain Laboratory., Angelaki, D., Benson, B. et al. A brain-wide map of neural activity during complex behaviour. Nature 645, 177–191 (2025). https://doi.org/10.1038/s41586-025-09235-0
Image Credits: AI Generated
DOI: https://doi.org/10.1038/s41586-025-09235-0
Keywords: neural decoding, choice representation, subcortical nuclei, cortex, brainstem, decision making, rodent neuroscience, population dynamics, electrophysiology, motor preparation
Tags: advanced population decoding techniquesbehavior and brain connectivitybrain-wide neural activity mapchoice signal dynamicscortical and subcortical brain regionsdecision-making in neurosciencefrontoparietal regions activitygroundbreaking neuroscience researchhindbrain structures in decision-makingneural representation of choiceprimate cortex versus rodent studiessingle-cell neural analyses
