Summary: By performing wide-field two-photon calcium imaging in mice performing a temporal expectation task, the researchers simultaneous recorded thousands of neurons across the secondary motor cortex (M2) and the posterior parietal cortex (PPC).
The team discovered that these regions are neither completely locked in sync nor entirely independent. Instead, they operate via a balanced dynamic where sparse inter-regional connections provide enough coupling to maintain overall temporal alignment, while widespread neural fluctuations allow each region to preserve its own local temporal flexible representation.
Key Facts
- Sequential Population Dynamics: Both the secondary motor cortex (M2) and posterior parietal cortex (PPC) track elapsed time through sequential patterns of neuronal activation, with distinct neuronal ensembles firing at successive temporal intervals.
- Dual Error Types Uncovered: Quantitative decoding revealed two distinct timing error modes during interval prediction:
- Correlated Drift: Trials where both M2 and PPC misjudged elapsed time identically (systemic alignment).
- Independent Drift: Trials where one region drifted off-clock while the other maintained accurate timing (regional autonomy).
- The Sparse Coupling Mechanism: Using twin-connected recurrent neural networks (RNNs), the team proved that sparse inter-regional connectivity acts as a gentle tether, ensuring regions stay broadly synchronized without forcing them into rigid, identical timekeeping states.
- Widespread Noise Preserves Autonomy: Global neural activity fluctuations act as a stabilizing counterweight, preventing hyper-synchronization and allowing local cortical networks to independently track distinct streams of information when tasks demand flexibility.
- Broad Technological Impact: Beyond clarifying neurological disorders linked to disrupted inter-regional communication (such as schizophrenia or autism), this framework provides architectural blueprints for brain-inspired AI and multi-agent robotic control systems.
Source: Institute of Science Tokyo
The brain is constantly keeping track of time, even if one does not consciously notice it. This ability to sense elapsed time underlies many mental processes, such as movement, planning, working memory, decision-making, and learning.
Several brain regions can represent time in some way, especially areas in the frontal and parietal cortex involved in higher-order cognition. But if timing information shows up in many different areas of the brain at once, how do they stay in sync with each other?
Scientists have long wondered whether multiple brain regions all follow a single shared internal clock, or whether each region can keep its own local clock when needed. Both possibilities could be useful; sometimes brain regions must stay aligned to a single event, while in other situations they may need to track different streams of information independently. However, testing this idea has proven difficult, as it requires recording the activity of large numbers of neurons at the same time while animals perform well-designed tasks.
To tackle this knowledge gap, a research team led by Associate Professor Riichiro Hira from the Department of Physiology and Cell Biology, Institute of Science Tokyo (Science Tokyo), Japan, investigated how two brain regions coordinate their sense of time.
In their study, published online in Volume 17 of the journal Nature Communications on June 11, 2026, the researchers examined the secondary motor cortex and posterior parietal cortex in mice during an innovative time-tracking task.
The researchers first trained the mice to predict the timing of rewards as they alternated between 6-second and 12-second intervals. Over time, the mice learned this pattern and began to anticipate both the short and long reward intervals. The team then used wide-field two-photon calcium imaging to record the activity of thousands of neurons simultaneously in the two brain regions while the animals performed the task.
In both areas, the team observed sequential patterns of activity, with different neurons becoming active at different moments, indicating that both regions were representing elapsed time. Through decoding analyses, the researchers then estimated what point in time each brain region was representing on a given trial. This approach revealed two kinds of errors: cases where both regions misjudged time similarly and cases where only one region drifted off. This suggests these regions are neither perfectly synced nor fully independent.
To explain this phenomenon, the team built a computational model consisting of twin connected recurrent neural networks. By analyzing neural activity in this model, the researchers gained insights into the brain’s timekeeping mechanisms, as Hira explains, “We found sparse inter-regional connectivity promotes synchronization between the two regions, whereas widespread fluctuations prevent complete synchronization, allowing each region to preserve its own temporal representation.”
In other words, the brain may use a combination of weak coupling and global activity patterns to balance stability with flexibility across regions.
Taken together, these findings provide a new framework for understanding how the brain can share information across regions without forcing every area into the same exact state.
“In the future, our results may contribute to a better understanding of cognition, neurological disorders involving disrupted inter-regional coordination, and the development of brain-inspired artificial intelligence and robotic control systems that combine stability with flexibility,” concludes Hira.
Key Questions Answered:
A: It uses a hybrid approach. Rather than relying on a single master clock or completely isolated timers, the brain uses loosely connected local clocks across different cortical regions. This study shows that areas like the motor cortex and parietal cortex run their own sequential timing patterns, but share just enough sparse connections to keep their clocks aligned when needed.
A: Wide-field two-photon calcium imaging is an advanced neuroimaging technique that allows scientists to visually track the activity of thousands of individual brain cells simultaneously in real time across large surface areas of the brain. It was essential for this study because measuring timekeeping errors required watching how massive populations of neurons in two separate brain regions (M2 and PPC) fired at the exact same moment.
A: Traditional AI and robotic systems often struggle to balance unified control with localized flexibility; either all subsystems are strictly locked together (making them rigid and vulnerable to single points of failure) or they run entirely independently (making coordination difficult). The “sparse coupling” framework discovered in this study offers a blueprint for building AI architectures that remain synchronized on macro-goals while allowing individual sub-modules to process independent streams of data flexibly.
Editorial Notes:
- This article was edited by a Neuroscience News editor.
- Journal paper reviewed in full.
- Additional context added by our staff.
About this neuroscience research news
Author: Miki Yamaoka
Source: Institute of Science Tokyo
Contact: Miki Yamaoka – Institute of Science Tokyo
Image: The image is credited to Neuroscience News
Original Research: Open access.
“Independence and coherence in temporal sequence computation across the fronto-parietal network” by Hiroto Imamura, Fumiya Imamura, Reiko Hira, Yoshikazu Isomura & Riichiro Hira. Nature Communications
DOI:10.1038/s41467-026-73999-w
Abstract
Independence and coherence in temporal sequence computation across the fronto-parietal network
Time processing requires distributed and coordinated cortical dynamics, yet how multiple brain areas flexibly switch between coherent and independent temporal representations remains unclear.
Using mesoscale two-photon calcium imaging, we simultaneously recorded neuronal populations in the secondary motor cortex and posterior parietal cortex of mice performing a novel alternating-interval timing task. Both areas encoded elapsed time through similar high-dimensional sequential activity, and decoding analyses revealed both coherent temporal errors shared across areas and independent errors confined to one area.
Communication-subspace analysis showed that temporal information was distributed across multiple low-variance shared dimensions, whereas the dominant shared dimension preferentially encoded behaviour. A twin recurrent neural network model with sparse inter-network coupling and shared high-variance noise reproduced these experimental findings.
Perturbation and local Lyapunov exponent analyses further showed that different shared subspaces selectively promote coherent or independent modes. These results reveal how sparse coupling and shared global fluctuations enable robust yet flexible fronto-parietal temporal computation.
