Elsevier

Neuroscience

Volume 435, 21 May 2020, Pages 58-72
Neuroscience

Research Article
Transcranial Direct Current Stimulation of Supplementary Motor Region Impacts the Effectiveness of Interleaved and Repetitive Practice Schedules for Retention of Motor Skills

https://doi.org/10.1016/j.neuroscience.2020.03.043Get rights and content

Highlights

  • IP led to poor training performance but superior test performance compared to RP.

  • Consolidation within 6-h of finishing IP led to less forgetting than after RP.

  • Sleep-related consolidation after IP not RP, improved performance for the next 72-h.

  • Applying anodal tDCS at SMA during RP facilitated sleep-related consolidation.

Abstract

Interleaved rather than repetitive practice (RP) is associated with superior retention of motor skills. It has been argued that this results from improved post-practice consolidation reflected in greater offline gains following interleaved practice (IP). The magnitude of this offline benefit has been associated with greater recruitment of supplementary motor area (SMA) during encoding. The present study administered anodal or cathodal transcranial direct current stimulation (tDCS) during interleaved or RP respectively in an attempt to modify the activity at SMA and the concomitant retention outcomes commonly associated with these training formats. Sixty-nine participants were assigned to one of four experimental conditions that included: IP-sham, RP-sham, IP-cathodal tDCS, and RP-anodal tDCS. Real or sham stimulation at SMA was administered during practice of three unique 6-key discrete sequence production tasks which lasted approximately 20-min. Tests were administered prior to practice and immediately after practice as well at 6-h, 24-h, and 72-h after practice ended. As anticipated, IP resulted in poorer acquisition but superior offline gain. Enhanced offline gain following interleaved training resulted from rapid stabilization of performance within the first 6-h following encoding and overnight improvement that continued over multiple sleep episodes. Administration of anodal stimulation at SMA during RP improved performance during training compared to sham but this benefit was short lived as forgetting during the first 6-h after practice was consistent with that observed for the sham counterpart. However, supplementing RP with anodal stimulation at SMA did foster overnight offline performance gains not displayed by individuals that experienced RP in the absence of stimulation.

Introduction

Determining how motor skills are acquired and detailing the associated neuroplasticity resulting from extended practice has garnered extensive interest in recent years (Dayan and Cohen, 2011, Doyon et al., 2018, Doyon et al., 2009). While few question the importance of extensive practice for learning, how practice is organized must also be considered when trying to optimize skill acquisition. The contextual interference (CI) effect is a learning phenomenon that has emerged from addressing how best to schedule practice of multiple motor skills that need to be learned simultaneously (Shea and Morgan, 1979, Lage et al., 2015, Wright et al., 2016). Scheduling training of these motor skills so that greater CI is induced in practice leads to poor initial performance compared to a schedule that creates a lower level of CI while supporting superior long-term retention and transfer. In the context of learning motor skills, the effect of experiencing increased CI during practice has been examined for a wide variety of laboratory tasks (Shea and Morgan, 1979, Wright et al., 2005, Pauwels et al., 2014) as well as in applied situations (Goode and Magill, 1986, Smith and Davies, 1995). A robust retention benefit from exposure to a high CI practice environment has been revealed for a variety of subject populations (Porretta and O’brien, 1991, Del Rey, 2011, Sidaway et al., 2016), and has been used with some success in the clinical domain (Knock et al., 2000, Wambaugh et al., 2014).

Greater CI is often engineered during motor skill acquisition by scheduling the presentation of a number of to-be-learned motor tasks in an interleaved or random format. Interleaved practice (IP) is assumed to create relatively high interference throughout training because of the frequent change in task demands that are encountered by the learner across trials. Conversely, repetitive practice (RP) creates significantly less CI because it involves the repeated execution of the same motor task before facing subsequent practice with other motor tasks. IP as opposed to RP, that is practice with greater as opposed to less interference, slows early improvements during practice, referred to as online gains in performance, but supports superior long-term retention. Retention, in most studies, is assessed by administration of a set of test trials at a single time point often 24–72-h after the completion of practice. What is most striking during the test trials is that prior exposure to RP usually leads to significant forgetting. In contrast, individuals trained via IP exhibit stable or even enhanced performance, often called offline gain, when test trials are administered as much as 72-h after practice is over (Lin et al., 2011, Kim et al., 2016). These data suggest that IP may play an important role in facilitating post-practice consolidation, a process reported to be central to the establishment of a robust motor memory (McGaugh, 2000, Cross et al., 2007, Kim et al., 2016).

A common prediction emanating from early theoretical accounts (Lee and Magill, 1985, Shea and Zimny, 1988, Shea and Zimny, 1983) is that greater CI during practice is associated with significantly greater attention demand during IP compared to RP. Li and Wright (2000) who used a dual-task paradigm to examine the relative attention costs of IP and RP revealed that, during movement preparation, an additional dual-task cost existed for IP-trained individuals (Li and Wright, 2000). Some years later, the inclusion of neural imaging whilst engaging in IP and RP indicated that these practice environments differed not only in the overall cost of movement preparation, reflected in the increase in attention demand, but also in terms of the neural recruitment strategies that were adopted across training (Cross et al., 2007, Wymbs and Grafton, 2009, Lin et al., 2011). For example, Wymbs and Grafton (2009) noted that increasing the relative magnitude of CI during training was associated with greater recruitment of neural regions previously described as central to successful preparation and production of learned motor skills (Wymbs and Grafton, 2009, Dayan and Cohen, 2011, Doyon et al., 2018, Doyon et al., 2009). Furthermore, the relative contribution of many of these regions to movement preparation, including the primary motor cortex as well as both lateral (PMd, PMv) and medial premotor regions (pre-SMA, SMA), increased dramatically with additional IP but not RP. In contrast, RP was characterized by heightened activation of constituents of the default network implying less active preparation and increased task independent processing that likely contributes little to developing new motor memories.

The initial work by Grafton and colleagues (Cross et al., 2007, Wymbs and Grafton, 2009) was extended by Lin et al. (2011) who reported that the relatively greater activity at superior frontal gyrus during IP was highly associated with subsequent behavioral gains that occurred across a 72-h retention interval (Lin et al., 2011). As a result of this finding, Lin et al. (2012) used DLPFC and PMd as seeds to examine functional connectivity that was present at the time of test for these sites with other key motor planning regions and their relationship to the learning advantage afforded through IP. For younger adults, similar to those used in the present work, functional integration following three days of IP was only observed for DLPFC most prominently with SMA (Lin et al., 2012). Importantly, the magnitude of DLPFC-SMA connectivity determined the size of the learning advantage offered by IP beyond that observed from RP. Lin et al. (2012) proposed that the benefits of increased CI during skill acquisition for young adults is supervised by a “DLPFC executive network” constructed to oversee efficient retrieval of required motor chunks stored in SMA for successful implementation of a motor sequence (Sun et al., 2007).

On the basis of the aforementioned results, recent commentaries have highlighted the importance of the differential involvement of SMA during IP and RP in determining retention outcomes (Wright et al., 2016, Lin et al., 2018). The SMA is located in the superior frontal gyrus and consists of at least two distinct regions. The rostral component, the pre-SMA, projects to frontal regions whereas the more caudal part, the SMA-proper, targets numerous cortical sites including the primary motor cortex (Nachev et al., 2008). While the specific functions of the pre-SMA and SMA-proper remain far from clear, it is generally accepted that they contribute to a diverse set of planning processes including being central to sequence processing likely involving integration of primitive task elements into higher order representations (Kennerley et al., 2004, Nachev et al., 2008). This role for the SMA is noteworthy because separate behavioral work has revealed that IP but not RP supports the development of more resilient parsing of motor sequences, or motor chunks, which have been described as a behavioral signature for advanced motor sequence skill acquisition (Wright et al., 2005). This latter finding is consistent with the reported heightened involvement of SMA during high CI practice conditions.

Taken as a whole, these data implicate the SMA is a key neural “player” during motor skill training and the recruitment of this neural region across practice trials may account, at least in part, for the differential effectiveness of IP compared to RP formats for skill retention. The present experiment was designed to explicitly evaluate this proposal for SMA1 during IP and RP. To accomplish this goal, anodal and cathodal transcranial direct current stimulation (tDCS) was applied at SMA in order to modify the cortical activity at this region during RP and IP respectively. tDCS involves the passage of weak direct current between two electrodes placed on a participant’s head. The applied current flows between a positively charged anode and a negatively charged cathode. Since tDCS induces an intracerebral current flow, neuronal excitability of the targeted brain area can be modified in a polarity-specific manner. Generally, anodal stimulation (with reference to the target area) increases cortical excitability while cathodal stimulation has been reported to decrease excitability at M1 (Nitsche and Paulus, 2000, Reis and Fritsch, 2011).

tDCS has frequently been used to modulate M1 activity to influence motor learning (Ammann et al., 2016, Buch et al., 2017). However, there are only a few examples of this form of non-invasive neuro-modulation being utilized to influence activity at SMA with an intent to modify behavior (Vollmann et al., 2013, Hupfeld et al., 2017). Carter et al. (2017) explored the importance of SMA for maintaining stable bi-manual coordination (Carter et al., 2017) while Hupfeld et al. (2017) recently demonstrated a general improvement in movement preparation for a number of skills varying in complexity when practiced in the presence of anodal tDCS at SMA (Hupfeld et al., 2017). Work of Vollmann et al. (2013) proposed that SMA-proper but not pre-SMA is the primary contributor to visuomotor learning, the type of learning targeted in the present work. It should be noted however that the learning described by Vollmann et al. (2013) is akin to what is referred to as online performance described in the present work and elsewhere and is only part of the process involved in developing a new motor memory (Vollmann et al., 2013). The present work was an attempt to further examine the role of SMA for motor learning by modifying the extent of SMA activity during practice contexts that have been reported to involve either extensive (i.e., IP format) or limited (i.e., RP format) recruitment of SMA and to assess the accompanying memory development.

To accomplish this goal, participants were randomly assigned to one of four experimental conditions. Two sham control conditions (IP-Sham, RP-Sham) that did not involve real stimulation, but mimicked CI conditions typically used when addressing the impact of this practice feature on skill acquisition. Performance of set of motor sequences was evaluated during a period of practice (i.e., online) and a series of delayed tests (i.e., offline) using total response time (TT) as the primary dependent variable (Kim et al., 2018). It was expected that the individuals assigned to RP-sham would exhibit superior online performance compared to individuals exposed to IP-sham. It was further anticipated that RP participants would exhibit considerable forgetting across the retention interval whereas those individuals trained in an IP format would display stable or maybe enhanced offline performance manifest as relatively less loss in performance at each of the test time points (Shea and Morgan, 1979, Lage et al., 2015, Wright et al., 2016).

Two important novel conditions involved cathodal stimulation at SMA during IP (IP-CtDCS) or anodal stimulation at SMA during RP (RP-AtDCS). Based on numerous studies evaluating motor performance in the presence of tDCS, it was anticipated that for the IP-CtDCS condition the administration of cathodal stimulation in conjunction with IP would impede online performance. The disruption in SMA activity during training was also assumed to have consequences for subsequent offline performance which was predicted to be less stable manifest as relatively greater forgetting than observed for the IP-Sham condition. Alternatively, given RP-AtDCS was assumed to entail an increase activation at SMA throughout practice, improvements in online performance should be observed. Of greatest interest was if the heightened activity of SMA during RP would influence the resultant offline gains compared to the RP-Sham group.

It is important to note that the goal of the present work was to evaluate if stimulating SMA could be used to systematically shift online and/or offline performance outcomes commonly associated with RP and IP. At this time the most compelling case was if we could improve (i.e., using anodal stimulation) the performance of individuals that encountered RP rather than just make them perform poorer (i.e., using cathodal stimulation). Likewise, we focused on impeding performance in IP rather than just making them better. The reciprocal conditions, at least at this point are less critical (i.e., making IP even better, and RP even worse). Moreover, while stimulation was expected to move performance in the direction commonly observed for the alternative practice condition (e.g., RP-AtDCS toward IP-Sham) it was impossible to predict at this time if the stimulation used would be sufficient (i.e., current, duration, etc.) to equate the performance of these conditions. Hence, the use of the unbalanced design in the present work and analyses that focused on each practice schedule (IP or RP) separately rather than comparing performance of all experimental conditions in a single analysis.

To specifically address the aforementioned hypotheses we focused on the impact of cathodal and anodal stimulation during IP and RP respectively for online gain which was defined as the change in TT between a baseline and post training set of test trials administered before and after an equivalent period of RP or IP. In addition, the present work expanded the investigation of the influence of CI during training of distinct components of offline gain. Specifically, offline gain was evaluated in terms of the initial development of the motor memory after a short time interval (6-h) that was dependent on time-dependent consolidation processes evaluated as a change in TT from the post training test trials to an additional set of test trials conducted 6-h later (Abe et al., 2011). In addition, the influence of CI and the associated stimulation on long-term retention was assessed more extensively through the inclusion of multiple test trials administered over the 72-h time period following the completion of practice during which additional training and stimulation were absent (Abe et al., 2011, Buch et al., 2017).

Section snippets

Participants

Participants were right-handed undergraduate students (N = 69) that received course credit for their participation. One individual assigned to the RP-AtDCS failed to follow instructions and was removed from all reported analyses. Individuals had no prior experience with the experimental tasks and were unaware of the specific purpose of the study. All participants completed an informed consent approved by an Institutional Review Board before any involvement in the experiment.

Discrete sequence production task

The motor skill

Baseline DSP task performance

Total time data from the baseline test trials prior to any practice for all individuals was subjected to a 4 (Group: RP-Sham, IP-Sham, RP-AtDCS, IP-CtDCS) between-subject analysis of variance (ANOVA) which failed to reveal a significant main effect of Group, F(3,64) = 1.97, p = 0.13, ηp2 = 0.08. Thus, as expected, TT did not differ as a function of Group prior to any exposure to training and stimulation (RP-Sham, M = 3129 ms, SEM = 109 ms; IP-Sham, M = 2810 ms, SEM = 110 ms; RP-AtDCS, M

Increased CI during training fosters both time-dependent and sleep-related consolidation

Interest remains high in determining how practice can be optimized to facilitate learning. Increasing the extent of CI to which a learner is exposed is one approach used to enhance the retention of motor skills (Shea and Morgan, 1979, Lage et al., 2015, Wright et al., 2016). IP is frequently adopted to manufacture greater CI and has been associated with robust delayed retention benefits compared to a RP format that has been proposed to induce less CI (Lin et al., 2012, Lin et al., 2011). The

Limitations

The present work was an initial attempt to use tDCS in two novel practice contexts to evaluate the contribution of SMA to the well-documented long-term retention advantage associated with IP of multiple novel motor skills. This work wasn’t without limitations, two of which are particularly important to consider for future efforts to verify and advance the findings from this work. A critical assumption made in this experiment was that activity at SMA was changed when tDCS, either anodal or

References (68)

  • V. Jurcak et al.

    10/20, 10/10, and 10/5 systems revisited: their validity as relative head-surface-based positioning systems

    Neuroimage

    (2007)
  • M. Abe et al.

    Reward improves long-term retention of a motor memory through induction of offline memory gains

    Curr Biol

    (2011)
  • E.L. Abrahamse et al.

    Control of automated behavior: insights from the discrete sequence production task

    Front Hum Neurosci

    (2013)
  • U. Amadi et al.

    The homeostatic interaction between anodal transcranial direct current stimulation and motor learning in humans is related to GABA<inf>A</inf> activity

    Brain Stimul

    (2015)
  • C. Ammann et al.

    Modulating motor learning through transcranial direct-current stimulation: an integrative view

    Front Psychol

    (2016)
  • G. Batsikadze et al.

    Partially non-linear stimulation intensity-dependent effects of direct current stimulation on motor cortex excitability in humans

    J Physiol

    (2013)
  • R. Bottary et al.

    Insufficient chunk concatenation may underlie changes in sleep-dependent consolidation of motor sequence learning in older adults

    Learn Mem

    (2016)
  • R.M. Brown et al.

    Off-line processing: reciprocal interactions between declarative and procedural memories

    J Neurosci

    (2007)
  • E.R. Buch et al.

    Effects of tDCS on motor learning and memory formation: a consensus and critical position paper

    Clin Neurophysiol

    (2017)
  • M.J. Carter et al.

    Intentional switches between coordination patterns are faster following anodal-tDCS applied over the supplementary motor area

    Brain Stimul

    (2017)
  • N. Censor et al.

    Modification of existing human motor memories is enabled by primary cortical processing during memory reactivation

    Curr Biol

    (2010)
  • G. Cona et al.

    TMS of supplementary motor area (SMA) facilitates mental rotation performance: evidence for sequence processing in SMA

    Neuroimage

    (2017)
  • E.S. Cross et al.

    Neural substrates of contextual interference during motor learning support a model of active preparation

    J Cogn Neurosci

    (2007)
  • E. Dayan et al.

    Neuroplasticity subserving motor skill learning

    Neuron

    (2011)
  • P. Del Rey

    Effects of contextual interference on the memory of older females differing in levels of physical activity

    Percept Mot skills

    (2011)
  • J. Doyon et al.

    Contributions of the basal ganglia and functionally related brain structures to motor learning

    Behav Brain Res

    (2009)
  • J. Doyon et al.

    Current issues related to motor sequence learning in humans

    Curr Opin Behav Sci

    (2018)
  • S. Goode et al.

    Contextual interference effects in learning three badminton serves

    Res Q Exerc Sport

    (1986)
  • K.E. Hupfeld et al.

    Transcranial direct current stimulation (tDCS) to the supplementary motor area (SMA) influences performance on motor tasks

    Exp Brain Res

    (2017)
  • C.H. Janice Lin et al.

    Interleaved practice enhances skill learning and the functional connectivity of fronto-parietal networks

    Brain Mapp Hum

    (2013)
  • J.S. Jo et al.

    The protective effects of acute cardiovascular exercise on the interference of procedural memory

    Res Psychol

    (2018)
  • S.W. Kennerley et al.

    Organization of action sequences and the role of the pre-SMA

    J Neurophysiol

    (2004)
  • T. Kim et al.

    Improving novel motor learning through prior high contextual interference training

    Acta Psychol (Amst)

    (2018)
  • T. Kim et al.

    Allowing time to consolidate knowledge gained through random practice facilitates later novel motor sequence acquisition

    Acta Psychol (Amst)

    (2016)
  • T.R. Knock et al.

    Influence of order of stimulus presentation on speech motor learning: a principled approach to treatment for apraxia of speech

    Aphasiology

    (2000)
  • K. Kuriyama et al.

    Sleep-dependent learning and motor-skill complexity

    Mem Learn

    (2004)
  • G.M. Lage et al.

    Repetition and variation in motor practice: a review of neural correlates

    Biobehav Rev Neurosci

    (2015)
  • T.D. Lee et al.

    Can forgetting facilitate skill acquisition?

    Psychol Adv

    (1985)
  • Y. Li et al.

    An assessment of the attention demands during random- and blocked-practice schedules

    Q J Exp Psychol Sect A Hum Exp Psychol

    (2000)
  • C.H. Lin (Janice) et al.

    Contextual interference enhances motor learning through increased resting brain connectivity during memory consolidation

    Neuroimage

    (2018)
  • C.H.J. Lin et al.

    Age related differences in the neural substrates of motor sequence learning after interleaved and repetitive practice

    Neuroimage

    (2012)
  • C.H.J. Lin et al.

    Brain-behavior correlates of optimizing learning through interleaved practice

    Neuroimage

    (2011)
  • K. Matsunaga et al.

    Increased corticospinal excitability after 5 Hz rTMS over the human supplementary motor area

    J Physiol

    (2005)
  • M.A. Mayka et al.

    Three-dimensional locations and boundaries of motor and premotor cortices as defined by functional brain imaging: a meta-analysis

    Neuroimage

    (2006)
  • Cited by (14)

    • Improving consolidation by applying anodal transcranial direct current stimulation at primary motor cortex during repetitive practice

      2021, Neurobiology of Learning and Memory
      Citation Excerpt :

      It’s entirely possible that the stimulation used by Kim and Wright (2020), targeting SMA, inadvertently increased activation at PMd independent of or in conjunction with heightened activation at the SMA. Indeed, this possibility was noted by Kim and Wright (2020) on the basis of the size of the electrodes used and the positioning of the montage on the skull. If this was the case, it’s conceivable that the enhancement reported herein for individuals receiving anodal tDCS at M1 might have resulted from strengthening PMd-M1 connectivity as opposed to emerging from an independent effect occurring at M1.

    View all citing articles on Scopus
    View full text