Switching between internal and external modes: A multiscale learning principle

Abstract

Brains construct internal models that support perception, prediction, and action in the external world. Individual circuits within a brain also learn internal models of the local world of input they receive, in order to facilitate efficient and robust representation. How are these internal models learned? We propose that learning is facilitated by continual switching between internally biased and externally biased modes of processing. We review computational evidence that this mode-switching can produce an error signal to drive learning. We then consider empirical evidence for the instantiation of mode-switching in diverse neural systems, ranging from subsecond fluctuations in the hippocampus to wake-sleep alternations across the whole brain. We hypothesize that these internal/external switching processes, which occur at multiple scales, can drive learning at each scale. This framework predicts that (a) slower mode-switching should be associated with learning of more temporally extended input features and (b) disruption of switching should impair the integration of new information with prior information.

Anticorrelated inter-network electrophysiological activity varies dynamically with attentional performance and behavioral states

“…The fMRI-based finding of anticorrelated networks during wakeful rest has led to the notion that there is an intrinsic, state-independent, antagonistic relationship between the DMN and other networks (Fox et al, 2005). Under this framework, the brain may continuously shift between states that draw, respectively, from internally-and externally-oriented sources of information (Buckner et al, 2008, Honey et al, 2018. However, the concept has remained controversial, in large part due to technical limitations of fMRI (Murphy and Fox, 2017).…”

“…In addition to task-dependent activity, fMRI studies during wakeful rest have shown spontaneous anticorrelated DMN-DAN/SN activity in infraslow (<0.1 Hz) fluctuations of blood-oxygen-leveldependent (BOLD) signals (Fox et al, 2005, Fransson, 2005) -a finding that has remained contentious (Murphy and Fox, 2017). Persistence of such anticorrelated activity in task-free states would potentially suggest that functionally competing systems, characterized by continual switching between internally-and externally-biased modes of attention, are an intrinsic property of the brain (Buckner et al, 2013, Honey et al, 2018.…”

Preparation for upcoming attentional states in the hippocampus and medial prefrontal cortex

“…This is particularly interesting to examine in cases like the current study, where both external attention and memory retrieval are needed for the effective guidance of behavior. One hypothesis is that the hippocampus might rapidly fluctuate between internal and external modes, prioritizing either attention/encoding or memory retrieval at different timepoints (Hasselmo 1995;Hasselmo and Fehlau 2001;Hasselmo and Schnell 1994;Hasselmo, Wyble, and Wallenstein 1996;Honey, Newman, and Schapiro 2017;Meeter, Murre, and Talamini 2004;Patil and Duncan 2018;Tarder-Stoll et al 2019). Although there are "background" fluctuations between external and internal attention in the hippocampus, top-down goals or external factors (e.g., surprise) can also affect these fluctuations (Sinclair & Barense, 2019).…”

Correlated activity favors synergistic processing in local cortical networksin vitroat synaptically-relevant timescales

“…The view that correlated neural activity is favorable for neural information processing is widely held within the cognitive rhythms community and is based on the idea that correlation facilitates both communication between circuits and orchestration of processing within circuits. Correlated neural activity is understood to generate the coherent rhythms that are observed in local field potentials, electrocorticography, and electroencephalography which have been theorized to subserve specific computational or cognitive mechanisms (e.g., Hasselmo et al, 2002;Ward 2003;Norman et al, 2006;Lisman & Jensen, 2013;Newman et al, 2014;Fries, 2015;Honey et al, 2017). A dominant and relevant theme among such theories is the importance of synchrony, a form of correlated activity, for enabling and organizing information transmission in cortical circuits.…”