Unravelling common mechanisms of time perception and working memory by neurocomputational modelling

• Applicant: Professor Dr. Joachim Hass
• Subject Area: Cognitive, Systems and Behavioural Neurobiology ; Experimental and Theoretical Network Neuroscience
• Term: since 2026
• Project identifier: Deutsche Forschungsgemeinschaft (DFG) - Project number 578686159

Time perception and working memory are among the most fundamental cognitive abilities in many animals and in men, allowing us to maintain thoughts or sensory information over a period of time and to perceive and interact with the flow of time in an ever-changing world. Converging evidence from neuroscience, medicine and psychology suggests a strong interrelation between time perception and working memory. Despite this wealth of experimental data, no attempt has been made so far to explain the mechanisms between the interaction of both abilities using biophysically detailed neurocomputational modelling, which has the potential to integrate the data from the different disciplines. In particular, there are striking parallels between the currently most discussed models of both functions. During the first funding period, we have contributed to the understanding of the intensely debated mechanistic basis of both time perception and working memory. We have tested four prominent models of time perception within the framework of a detailed model of the prefrontal cortex (PFC) in terms of their ability to reproduce a range of prominent experimental findings. We found that the state-dependent network model and the ramping activity model were compatible with all data, while the striatal beat model was found too sensitive against noise and testing the synfire chain model exceeded the available computational resources. Furthermore, we assessed whether the most prominent neurocomputational working memory model, i.e., persistent activity in cell assemblies, could be implemented in the same PFC network model. Persistent activity could only be produced if the strong, naturally occurring heterogeneities in the inhibitory feedback to the cell assemblies are diminished, e.g., by a homeostatic plasticity rule. In the second funding period, we will now construct a neurocomputational model of both time perception and working memory within the same PFC network and thus, fill the knowledge gap about the neural mechanisms underlying the interactions between the two functions. We hypothesize that time perception and working memory compete for neural resources within the same network: Time perception requires a broad range of neural and synaptic parameters to encode different interval durations, while working memory relies on largely homogeneous cell assemblies. This conflict may be the basis for the frequently reported decrease in performance in both tasks when they are performed at the same time, as well as the recently found influence of manipulated subjective duration on working memory performance. The prediction of the model on the latter paradigm will be tested in a human experiment with electroencephalographic (EEG) recordings. Taken together, this project has the potential to provide, for the first time, a coherent neurocomputational modelling framework for the interaction of these essential cognitive abilities.

• DFG Programme: Research Grants