Action Prediction Allows Hypothesis Testing via Internal Forward Models at 6 Months of Age

Abstract

We propose that action prediction provides a cornerstone in a learning process known as internal forward models. According to this suggestion infants’ predictions (looking to the mouth of someone moving a spoon upward) will moments later be validated or proven false (spoon was in fact directed toward a bowl), information that is directly perceived as the distance between the predicted and actual goal. Using an individual difference approach we demonstrate that action prediction correlates with the tendency to react with surprise when social interactions are not acted out as expected (action evaluation). This association is demonstrated across tasks and in a large sample (n = 118) at 6 months of age. These results provide the first indication that infants might rely on internal forward models to structure their social world. Additional analysis, consistent with prior work and assumptions from embodied cognition, demonstrates that the latency of infants’ action predictions correlate with the infant’s own manual proficiency.

The social foundation of executive function

“…Following Gredebäck, Lindskog, et al (), this is a combined measure of infants' action prediction and action evaluation scores ( r = .41, p < .000). The measure was constructed by first reversing the action prediction variable and then averaging standardized scores from the two tasks.…”

“…The measure was constructed by first reversing the action prediction variable and then averaging standardized scores from the two tasks. The action prediction and action evaluation tasks are described briefly below (see Gredebäck, Lindskog, et al, for detailed task information). Positive values on this measure indicate that the participant was faster at predicting goal‐related actions (i.e., action prediction) and/or reacted with more surprise when events unfolded in an inappropriate way (i.e., action evaluation).…”

“…Rudimentary cognitive abilities for processing social information are present shortly after birth (Farroni, Csibra, Simion, & Johnson, ; Goren, Sarty, & Wu, ). Important social cognitive skills, such as the ability to form internal models (i.e., the capability to predict and interpret other peoples' actions based on previous experiences; see Gredebäck, Lindskog, Juvrud, Green, & Marciszko, ) and gaze following (i.e., the ability to share attention with others based on their gaze direction) are evident before 6 months of age (Butterworth & Jarret, ; Gredebäck, Lindskog, et al, ). These abilities are critical for successful social interactions (Southgate & Vernetti, ) and transfer of social knowledge (Morales et al, ).…”

Six‐Month‐Old Infants Predict Agents’ Goal‐Directed Actions on Occluded Objects

“…This work suggests that infants in previous studies using looking time measures may not have simply been responding passively to the action taken by the agent, but may instead have actively generated a prediction about her behavior and responded when that prediction was violated. However, at 6 months, the ability to generate action predictions may still be developing; Gredebäck, Lindskog, Juvrud, Green, and Marciszko () found individual differences in 6‐month‐olds’ ability to generate action predictions, and further found that this ability correlates with infants’ ability to evaluate whether actions unfold appropriately (e.g., an object placed in an agent's outstretched hand rather than on her head), suggesting an emerging competence for action understanding at 6 months. Indeed, action prediction may be supported at least in part by the motor system (Gredebäck & Falck‐Ytter, ; Krogh‐Jespersen & Woodward, ; Kanakogi & Itakura, ), which at 6 months is still developing.…”

“…We chose to test 6‐month‐old infants, for two reasons. First, infants of this age have previously been shown to successfully infer agents’ goals from their preferential reaching behavior (Luo & Baillargeon, ; Woodward, ) and to make gaze predictions based on agents’ goals (Kim & Song, ; Kochukhova & Gredebäck, ), albeit with some variability (Gredebäck et al., ). Second, and most importantly, 6‐month‐olds’ ability to maintain representations of two occluded objects (in the absence of a social agent) is well understood (Káldy & Leslie, ; Kibbe & Leslie, , , in press).…”