Learning, remembering, and predicting how to use tools: Distributed neurocognitive mechanisms: Comment on Osiurak and Badets (2016)

Abstract

The reasoning-based approach championed by Francois Osiurak and Arnaud Badets (Osiurak & Badets, 2016) denies the existence of sensory-motor memories of tool use except in limited circumstances, and suggests instead that most tool use is subserved solely by online technical reasoning about tool properties. In this commentary, I highlight the strengths and limitations of the reasoning-based approach and review a number of lines of evidence that manipulation knowledge is in fact used in tool action tasks. In addition, I present a "two route" neurocognitive model of tool use called the "Two Action Systems Plus (2AS+)" framework that posits a complementary role for online and stored information and specifies the neurocognitive substrates of task-relevant action selection. This framework, unlike the reasoning based approach, has the potential to integrate the existing psychological and functional neuroanatomic data in the tool use domain. (PsycINFO Database Record

Figure 1

Graphic depiction of the Two Action Systems Plus (2AS+) model of tool use. A left-lateralized posterior temporal/inferior parietal system subserves storage of abstract, multimodal manipulation knowledge (blue), which is translated into sensorimotor representations enabling tool use production in a bilateral frontoparietal network, with additional “tuning” based on current visual and somatosensory input (purple). The portion of the network specialized for action selection (green) subserves a biasing signal from the inferior frontal cortex that aids in selection of potential actions from a temporary repository or “buffer” in the left supramarginal gyrus.

Figure 2

Schematic illustration of possible role of manipulation knowledge (ventro-dorsal stream) at several iterative stages of control and learning of tool actions. Manipulation knowledge consists of abstracted representations of prior viewed and/or performed tool use actions, including the visual appearance of the body and tools, abstract movement shape trajectory information, and somatosensory information, serving as target or goal states for feedback and feedforward control. Left: at Time 1 (T1) in a given episode, manipulation knowledge from prior experience is compared with the current state of the body and environment to generate a prediction of the motor command needed to achieve the target state. At Time 2 (T2), predicted sensory feedback is compared with the target state, and any mismatch (predicted error) is used to generate a final motor command aimed at minimizing the error. At Time 3 (T3), sensory feedback of the new body and environment states arising from the motor command adjusts the target state (learning).

Figure 3

Voxel-based lesion symptom mapping analysis showing regions in the supramarginal gyrus and anterior insula/inferior frontal gyrus significantly predicting poor performance on tools associated with distinct ‘use’ and ‘move’ actions (“conflict” tools) controlling for performance on “non-conflict” tools. Reproduced with permission from Watson & Buxbaum, 2015.

Figure 4

Examples of the visual stimuli used by Kalenine et al. (2013) to demonstrate that tool (e.g., kitchen timer) use representations are signfificantly more activated in visual contexts consistent with use (e.g., kitchen countertop) than move (e.g., kitchen drawer). Reproduced with permission from Kalenine et al., (2013).