Chunking as the result of an efficiency computation trade-off

Abstract

How to move efficiently is an optimal control problem, whose computational complexity grows exponentially with the horizon of the planned trajectory. Breaking a compound movement into a series of chunks, each planned over a shorter horizon can thus reduce the overall computational complexity and associated costs while limiting the achievable efficiency. This trade-off suggests a cost-effective learning strategy: to learn new movements we should start with many short chunks (to limit the cost of computation). As practice reduces the impediments to more complex computation, the chunking structure should evolve to allow progressively more efficient movements (to maximize efficiency). Here we show that monkeys learning a reaching sequence over an extended period of time adopt this strategy by performing movements that can be described as locally optimal trajectories. Chunking can thus be understood as a cost-effective strategy for producing and learning efficient movements.

A Network Neuroscience of Human Learning: Potential to Inform Quantitative Theories of Brain and Behavior

“…While inference is a useful tool in learning, so too is chunking, or the ability to combine fine-scale solutions into bigger sets of solutions. For example, one can combine a series of motor responses into a single response [140,152], thereby reducing the complexity of the required input–output mapping.…”

A model of flexible motor sequencing through thalamic control of cortical dynamics

“…Animals have the remarkable ability of flexibly performing long and complex sequences of movements [42,13,52,28]. In humans, dance provides an illustration of this ability.…”

Parcellation of motor sequence representations in the human neocortex

“…These motor chunks could also be further organized into larger chunks or flexibly re-used to create different sequences (Sakai et al, 2003). Such hierarchical structure would greatly reduce computational cost for planning and executing long 30 motor sequences (Ramkumar et al, 2016). An alternative idea is that sequences are non-hierarchically represented as a continuous set of transition probabilities between neighbouring movements (Hunt and Aslin, 2001;Reber, 1967;Stadler, 1992;Verwey and Abrahamse, 2012).…”

“…Rather than analysing the increases or decreases of spatially smoothed activity, representational fMRI analysis makes inferences based on the similarity (or dissimilarity) of multivariate activity patterns across multiple experimental conditions (Ban and 60 Welchman, 2015;Chikazoe et al, 2014;Ejaz et al, 2015;Kriegeskorte et al, 2008;Yokoi et al, 2018). One potential problem in applying this method to study the hierarchical organization of movement sequences is that the specific organisation of motor memories may be different across individuals (Jimenez et al, 2011;Ramkumar et al, 2016), and is often influenced (and hence confounded) by biomechanical 65 constraints (Koch and Hoffmann, 2000). To address this problem, we first established a new behavioural paradigm that allowed us to experimentally manipulate the structure of motor sequence representations.…”