Edited by: Björn H. Schott, Leibniz Institute for Neurobiology (LG), Germany
Reviewed by: Max-Philipp Stenner, Otto-von-Guericke University Magdeburg, Germany; Leibniz Institute for Neurobiology, Germany; Patricia A. Shewokis, Drexel University, United States
This is an open-access article distributed under the terms of the Creative Commons Attribution License (CC BY). The use, distribution or reproduction in other forums is permitted, provided the original author(s) and the copyright owner(s) are credited and that the original publication in this journal is cited, in accordance with accepted academic practice. No use, distribution or reproduction is permitted which does not comply with these terms.
Positive effects of variable practice conditions on subsequent motor memory consolidation and generalization are widely accepted and described as the contextual interference effect (CIE). However, the general benefits of CIE are low and these benefits might even depend on decreased retest performances in the blocked-practicing control group, caused by retroactive inhibition. The aim of this study was to investigate if CIE represents a true learning phenomenon or possibly reflects confounding effects of retroactive inhibition. We tested 48 healthy human participants adapting their reaching movements to three different force field magnitudes. Subjects practiced the force fields in either a Blocked (B), Random (R), or Constant (C) schedule. In addition, subjects of the Blocked group performed either a retest schedule that did (Blocked-Matched; BM) or did not (Blocked-Unmatched; BU) control for retroactive inhibition. Results showed that retroactive inhibition did not affect the results of the BU group much and that the Random group showed a better consolidation performance compared to both Blocked groups. However, compared to the Constant group, the Random group showed only slight benefits in its memory consolidation of the mean performance across all force field magnitudes and no benefits in absolute performance values. This indicates that CIE reflects a true motor learning phenomenon, which is independent of retroactive inhibition. However, random practice is not always beneficial over constant practice.
It is widely accepted that variable practice conditions can be beneficial for motor memory consolidation (Schmidt,
Although CIE seems to be robust, there is no widely accepted hypothesis that accounts for this effect. Classical explanations include the elaboration hypothesis (Shea and Zimny,
So far, CIE and retroactive inhibition were discussed in the context of skill learning, in which most CIE studies were conducted. Skill learning is commonly defined as a “set of processes associated with practice or experience leading to relatively permanent changes in the capacity for skilled movement” (Schmidt et al.,
Therefore, the first purpose of this study is to control for the confounding effects of retroactive inhibition and examine the validity of the CIE in force field adaptation. The second purpose of this study is to examine if variable practice schedules (blocked and random) outperform a constant practice schedule even if subjects of the constant group have the advantage of adapting their reaching movements only to a single force field.
This study tested 48 healthy right-handed participants (24 ± 4 years; 10 women) with no previous experience at a robotic manipulandum. Handedness was tested by the Edinburgh inventory (Oldfield,
The experimental task was implemented at a robotic manipulandum (Kinarm End-Point Lab, BKIN Technologies, Kingston, ON, Canada) which can produce forces
Apparatus and task.
We will briefly describe the experimental task, which can be found elsewhere in more detail (Thürer et al.,
To provide similar movement times across trials and subjects, visual feedback was implemented in every single trial. The feedback was given
To induce motor adaptation and subsequent memory consolidation, we implemented velocity-dependent counter-clockwise directed force fields at the robotic manipulandum. These force fields perturbed the participants’ movements and typically degraded their initial motor performance leading to curved hand trajectories (
Participants were equally distributed into four groups (Blocked-Matched, BM; Blocked-Unmatched, BU; Constant, C; Random, R; each
On day 1, all participants received instructions about the behavioral task and performed 144 familiarization trials under null field conditions (motors of the robot were turned off) with two breaks of 30 s after every 48th trials. Then, participants performed a baseline measurement consisting of 30 null field trials. After that, all participants performed 540 force field trials during Practice, with a different force field schedule according to their group allocation. To avoid fatigue, participants had a 30 s break after each 60th trial. The participants performed all trials on day 1 with their dominant right hand.
The practice schedule was identical between the two Blocked groups (BM, BU) but different for the Random and Constant groups. Participants of the Blocked groups performed the three force field magnitudes (8, 15, 22 Ns/m) in a blocked order. Therefore, all trials of one specific magnitude were practiced first, before switching to the next magnitude. This resulted in three blocks, each containing 180 trials of one specific force field magnitude. The Random group performed a highly-variable practice schedule so that the three force field magnitudes changed on a single-trial level. For the Constant group, each participant practiced only one specific force field magnitude (e.g., 15 Ns/m) and, thus, encountered no force field variability at all. The force field magnitude (for C) and the magnitude order (for BM, BU, and R) was counter-balanced across participants so that the mean force field magnitude was 15 Ns/m on the group level (see
On day 2, all participants performed a Posttest and Transfer test. To quantify Posttest performance, all participants performed 18 force field trials divided into three blocks with each block representing one force field magnitude. Then, participants performed 60 trials of a constant force field magnitude with their non-dominant left hand (Transfer test) to investigate long-term effects on the contralateral hand indicated by a previous study from our group (Thürer et al.,
The order of force field magnitudes on day 2 differed between groups. For the BM group, the magnitudes in Posttest were in a reversed order compared to Practice and, thus, “matched” in terms of a reduced effect of retroactive inhibition on the first block of the retest schedule (
For the statistical analyses, mean performance for the first and the last six trials of the Practice session (Practice FT, Practice LT) was computed. Posttest performance was computed by the mean of the first, middle, and last six Posttest trials (Posttest FT, MT, LT) and the mean across all 18 Posttest trials (Posttest ALL). Contralateral Transfer performance was quantified by the initial six Transfer trials (Transfer FT) and the whole 60 Transfer trials (Transfer ALL).
To test for the possible influence of retroactive inhibition on CIE, we performed mixed-model 2*2 ANOVAs with the factors time (Practice LT, Posttest FT; Practice LT, Posttest ALL) and group (BM, BU). For a possible effect on the generalization from one hand to the other, the factor time was adjusted accordingly (Practice LT, Transfer FT; Practice LT, Transfer ALL). In addition, we investigated if random practice even outperforms constant practice by using standard Fischer
It is widely accepted that
All parameters were tested for normal distribution and homogeneity of variances using Shapiro-Wilk and Levene test. For two independent group comparisons no normal distribution was given and Mann-Whitney-
The progress in motor performance for both Blocked groups is depicted in
Descriptive results.
Consolidation of motor memory (from Practice to Posttest) did not differ between Blocked groups regarding their recall of the first force field magnitude (
Deeper investigation of the behavioral results.
We further investigated if retroactive inhibition affected the generalization from the dominant (Practice) to the non-dominant (Transfer) hand. No differences between Blocked groups [
The second aim of this study was to examine if constant practice leads to better memory consolidation of only one force field magnitude than random practice and if random practice outperforms constant practice in the recall of multiple force field magnitudes. The progress in motor performance for the Random and the Constant group is depicted in
However, it is important to mention here that this consolidation effect occurred due to performance differences at the end of Practice, for there is no group difference regarding absolute Posttest values [Posttest FT:
Our results showed no differences between the two Blocked groups although the Posttest schedule of one group (BM) did and the other schedule (BU) did not control for retroactive inhibition. Compared to the Random group both Blocked groups showed a limited memory consolidation, which depicts that retroactive inhibition does not account for CIE in motor adaptation tasks. Comparisons between Random and Constant groups showed a similar memory consolidation for each single force field magnitude. However, the Random group outperformed the Constant group in its mean memory consolidation across all three force field magnitudes.
The experimental procedure of the BM group controlled for possible confounding effects of retroactive inhibition within the first Posttest trials. Nevertheless, BM performed similar to BU and its memory consolidation was hampered compared to the Random group. These findings contradict previous skill learning studies (Poto,
These benefits for the Random group were not observed when testing for the generalization of memory to the contralateral hand. Our results did not report an enhanced generalization from the Practice on the right hand to the Transfer on the left hand. This finding contradicts in some extent with the literature, which frequently showed CIE for transfer tests in skill learning tasks (e.g., Shea and Morgan,
Taken these results together, it seems that retroactive inhibition is not able to explain the frequently observed phenomenon of contextual interference. Therefore, other approaches like the elaboration (Shea and Zimny,
Although CIE reflects a widely accepted phenomenon and seems to be unaffected by retroactive inhibition in motor adaptation tasks, it is not clear whether random practice is always beneficial over constant practice. Our results showed that the benefits of random compared to constant practice regarding motor memory consolidation occur only if multiple force field magnitudes are retested. This indicates that memory consolidation of a single task might not be improved by a highly variable practice schedule. This concurs with the especial skill effect for skill learning (Breslin et al.,
These results concur with the assumption that uncertainty is taken into account for updating internal models (Tan et al.,
This finding is also in line with our correlation results. We were able to show that both, an increased motor error and an increased motor variability during Practice hamper the more blocked Posttest performance. Especially the absolute Posttest performance of the Random group was reduced by the confounding effect of motor variability. However, it is important to note that the absolute values of Posttest performance did not differ significantly between groups. Nevertheless, derived from a practical perspective, Random practice might be the better choice of scheduling a practice session since it leads to similar results than constant practice but has the opportunity to increase mean memory consolidation of multiple force field magnitudes.
It is important to discuss the diverging motor performance of the Random group during Practice from the other groups. Due to the variable character of the practice schedule, the Random group’s motor performance is reduced. This raises the important question of whether CIE observed in this study is actually a benefit or the Random group during consolidation or just reflecting the disadvantage during Practice. Accordingly, Blocked, Constant, and Random groups might not deviate from each other in their Posttest and Transfer performance because their internal model was equally adapted after Practice. However, switching the force field from trial to trial in the Random group induces noise in the motor error parameter. It is hard to control for this possible confounding factor since both, adding constant trials at the end of the Practice session for the Random group or having counterbalanced random and blocked Posttest schedules, would come with other limitations. However, the previous CIE literature shows that absolute performance differences are also observed in force field adaptation studies, using a lower amount of force field variability (Thürer et al.,
This study showed some minor limitations, which we would like to address. The Constant group trained the same amount of trials as the other groups but each subject of only one force field magnitude. Therefore, this group was able to draw on a greater practice experience for one specific magnitude compared to the other groups. We cannot state how much this affected the results but from a practical perspective, it was important to have the same amount of practice time for each group.
The force field magnitudes might have been too different and, thus, induced a too high practice variability in the Random group. This might be the reason why we were not able to show absolute Posttest and Transfer test performance benefits for the Random group. In a previous study with a lower amount of variability, we were able to show these absolute benefits after Random practice in the transfer test on the contralateral hand (Thürer et al.,
The order of Post- and Transfer tests was not counter-balanced. Therefore, similar group performances in the first Transfer trials might be caused by the 18 Posttest trials. However, we were previously able to show that contralateral transfer from the dominant to the non-dominant hand after random practice is almost independent of the Posttest performance (Thürer et al.,
Generalization was quantified by comparing Practice trials on the right hand with Transfer trials on the left hand, without normalizing for left hand baseline performance. We believe that this issue has not affected our results much since we investigated the interaction effect between groups and time. A baseline performance would serve as a constant for the baseline correcting function for every single subject. As long as there is no group difference between the left hand baseline performance, which is highly unlikely, this should not have affected our results much.
In this study, we investigated motor adaptation and not skill acquisition and, therefore, our interpretations cannot be generalized to skill learning tasks. However, from a theoretical point of view, although skill learning involves different brain areas than motor adaptation (Debas et al.,
In this study, we were able to show that the CIE represents a valid learning phenomenon that is not affected by retroactive inhibition. Furthermore, we were able to show that the benefits of random practice are more related to the memory consolidation of multiple tasks/parameters. However, variability, in general, must not always be beneficial regarding a single task/parameter or regarding the absolute performance values in a posttest. However, it remains unsolved how the motor system uses variability to improve subsequent motor memory consolidation, which needs further investigation on the neurobiological level.
The study was approved by the Institutional Review Board. All participants were informed about the protocol and gave their written informed consent in accordance with the Declaration of Helsinki.
The study was designed by BT, AF, and TS. SG recorded the data and BT performed the data processing and statistical analysis. BT wrote the first draft of the manuscript and all authors contributed to manuscript revision. All authors read and approved the submitted version of the manuscript.
The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.
We like to thank Alexander Wolpert for his technical assistance and Ernst Hossner for the fruitful discussions and his valuable input. A preprint of this manuscript is available at bioRxiv (Thürer et al.,
The Supplementary Material for this article can be found online at: