Absence of post-learning motor activity effects on memory for motor-related words

ABSTRACT

Drawing a hypothesis from embodied theories of memory, van Dam, Rueschemeyer, Bekkering and Lindemann [(2013). Embodied grounding of memory: Toward the effects of motor execution on memory consolidation. The Quarterly Journal of Experimental Psychology, 66(12), 2310–2328] showed that recognition performance for action words could be modulated by actions performed during the retention interval, suggesting that motor actions during the retention interval affect memory consolidation. The results of 4 experiments from two different laboratories, designed to replicate and extend the van Dam et al. motor consolidation effect, are presented here. Two of the experiments (n = 30 and n = 44) exactly and independently replicated the experimental design and conditions of the original experiment. Yes/No recognition scores plus additional analysis of response times showed no motor consolidation effects. A third experiment (n = 44) manipulating type of processing during encoding also failed to find significant motor consolidation effects. Finally, a fourth experiment (n = 120) following a more standard reconsolidation paradigm, involving 24-hour intervals between learning and motor behaviour, and a 24-hour delayed test, also found null effects. The absence of effects of motor execution on memory consolidation is discussed in terms of the implications of these findings for the embodiment approach to cognition.

KEYWORDS:

• Embodiment
• action
• memory
• replication

Disclosure statement

No potential conflict of interest was reported by the author(s).

Notes

1 We opted for this device as it enabled the two types of relevant movements (pressing and twisting) in a single apparatus, and because it presented a close functional similarity to the set up described in the reference experiment (Van Dam et al., 2013), which involved two separate but barely described response devices.

2 Occasionally, hits or false alarms with values of 0 were found. Although there are different approaches to dealing with this problem (and there are no perfect solutions) in our data analyses we have used 2 approaches, following Stanislav and Teodorov (1999). First, we adjusted only the extreme values by replacing rates of 0 with 0.5/n and rates of 1 with (n−0.5)/n, where n is the number of signal (hit) or noise (false alarm) trials (Macmillan & Kaplan, 1985). We assume that this is a biased solution, because we are not treating the data points equally. And for that reason, additionally, we calculated also the non-parametric measure A' and conducted all the analyses, verifying the same pattern of results in all the experiments.

Additional information

Funding

The work reported here received financial support from by the Spanish Ministry of Economy and Competitiveness (grant PSI2017-82748-P), and by Consejería de Educación, Junta de Castilla y León (grant SA052G18). Address correspondence to María A. Alonso (maalonso@ ull.edu.es). The raw data for all the experiments in this study are available at the Open Science Framework (https://osf.io/bx945/) [use this link for peer reviewing: https://osf.io/bx945/?view_only=8fe41d5c9b1b4a7d8f60a75aa124948e]