Exp Brain Res (2010) 203:367–380

DOI 10.1007/s00221-010-2239-x

RESEARCH ARTICLE

Reach adaptation to explicit vs. implicit target error

Brendan D. Cameron · Ian M. Franks · J. Timothy Inglis · Romeo Chua

Received: 27 July 2009 / Accepted: 28 March 2010 / Published online: 11 April 2010© Springer-Verlag 2010

Abstract The adaptation of reaching movements has typi-cally been investigated by either distorting visual feedback ofthe reaching limb or by distorting the forces acting upon thereaching limb. Here, we investigate reach adaptation whenerror is created by systematically perturbing the target of thereach rather than the limb itself (Magescas and Prablanc in JCogn Neurosci 18: 75–83, 2006). SpeciWcally, we investigatehow adaptation is aVected by (1) the timing of the perturba-tion with respect to the movement of the eye and the handand (2) participant awareness of the perturbation. In Experi-ment 1, participants looked and pointed to a target that disap-peared either at the onset of their eye movement or shortlyafter their eye movement and then reappeared, displaced tothe right, at the completion of the reach. In Experiment 2, wemade the target displacement more explicit by leaving thetarget at its initial location until the end of the reach, at whichpoint it was displaced to the right. In Experiment 3, we extin-guished the target at the onset of the eye movement but alsoinformed participants about the presence and magnitude ofthe perturbation. In the no-feedback post-test phase, partici-pants for whom the target disappeared during the reach dem-onstrated much stronger aftereVects of the perturbation,misreaching to the right, whereas participants for whom thetarget stayed on until reach completion demonstrated rapidextinction of rightward misreaching. Furthermore, partici-pants who were informed about the target perturbation exhib-ited faster de-adaptation than those who were not. Our resultssuggest that adaptation to a target displacement is contingenton the explicitness of the target perturbation, whether this isachieved by manipulating stimulus timing or instruction.

Keywords Motor adaptation · Reach error · Saccade · Awareness

Our ability to learn new motor skills and to optimize theones we already have relies on detecting error in our actionsand modifying subsequent actions to compensate for thaterror. Consider, for instance, the actions of a novice astro-naut, encountering for the Wrst time, the altered limbdynamics created by wearing a spacesuit in a weightlessenvironment. Her Wrst reach will probably miss its target,but over repeated attempts, she will reduce and then elimi-nate reach error. When faced with novel reaching environ-ments, people are very eVective at adapting theirmovements: We adapt quickly to altered limb dynamics,such as when force perturbations are applied to the reachinglimb by a robotic arm (e.g., Shadmehr and Mussa-Ivaldi1994), and we adapt quickly to altered visual feedback,such as when prisms rotate our visual Weld and alter the cor-respondence between motor commands and visual feed-back (Welch 1986; Redding et al. 2005). The speed andease with which we are able to adapt to these conditionssuggests that the human motor system is highly Xexible andis constantly primed to detect and eliminate movementerror. How we process movement error, however, and theextent to which adaptation persists after a perturbation isremoved, may depend greatly on the extent to which weattribute the error to ourselves.

A key determinant of whether we attribute error to our-selves may be the perceived stability of the target of ourreach. If, for instance, a target visibly jumps to a new loca-tion while we are reaching for it, we might process thereach error diVerently than if the same amount of erroroccurred when we reached to a stationary target. In the Wrstcase, we would, presumably, attribute the error to the

B. D. Cameron · I. M. Franks · J. Timothy Inglis · R. Chua (&)School of Human Kinetics, University of British Columbia, 6081 University Blvd., Vancouver, BC V6T 1Z1, Canadae-mail: romeo.chua@ubc.ca

123

368 Exp Brain Res (2010) 203:367–380

behavior of the target (such as when a Xy jumps away whenwe attempt to swat it), while in the latter case, we mightattribute the error to ourselves—our visuomotor system mightconclude that it issued an incorrect command. Diedrichsenet al. (2005) have provided behavioral and neural evidencethat these two kinds of error are processed diVerently. Theyshowed that people adapt less when a target visibly jumpsat movement onset relative to when visual feedback ofthe hand is rotated. Similar amounts of reach error are pro-duced in the two conditions, but participants appear toadapt more when the trajectory of the reach, rather than theposition of the target, is perturbed. Diedrichsen et al. alsoshowed that neural activity diVered, to some extent, whenparticipants responded to these two types of error; when thetarget moved, greater activity was observed in the superiorparietal lobule and in the striatum.

The results of Diedrichsen et al. (2005) are consistent withthe idea that the presence of a visible target jump alters theway we attribute error and, consequently, how we learn fromthat error. One would predict, therefore, that if a target jumpis rendered invisible, reaching movements should adapt tothe reach error that the target jump produces. However, thestory appears to be somewhat more complicated: Unseen tar-get perturbations seem to produce adaptation if the targeterror is presented at the completion of the movement(Magescas and Prablanc 2006), but not if the target erroroccurs while the hand is in Xight (Magescas et al. 2009).

Magescas and Prablanc (2006) demonstrated adaptationto terminal target error in an experiment where they askedparticipants to look and point to a target that disappearedwith the onset of their eye movement. When the handlanded, the target reappeared at a location slightly to theright of its initial location. Over multiple trials, the size ofthe shift gradually increased, as did the amplitude of theparticipants’ reaching movements, even though the partici-pants never became aware of the location change. In thepost-test phase of the experiment, participants exhibitedvery strong aftereVects (i.e., when visual feedback wasremoved, they continued to aim far to the right of targetlocations). This contrasts with the eVect observed when thetarget was moved during the eye and hand movement,allowing participants to correct their reach online. In thiscase, no aftereVects were observed (Magescas et al. 2009),suggesting that an unseen target jump does not producereach adaptation when it occurs online.1 Note that Mages-cas et al.’s (2009) experiment is eVectively the ‘invisible’

equivalent of Diedrichsen et al.’s (2005) target jump condi-tion, in which participants corrected their movement onlineto a visible target jump. Taken together, these two studiesmight indicate that online target error, whether explicit (vis-ible target displacement) or implicit (invisible target dis-placement), prevents signiWcant adaptation of the reach.The goal of the present study was to examine the role ofperceived target stability in reach adaptation without theconfounding eVects of online error correction.

We attempted to manipulate perceived stability of thereach target in two ways: (1) by varying the timing of thetarget perturbation with respect to the eye and the hand and(2) by informing participants about the presence of the per-turbation. We examined reach adaptation in four groups ofparticipants, all of whom were directed to look and point toa target. For the Wrst group (SacStart), we extinguished thetarget at the onset of the eye movement and re-illuminatedit at hand touchdown, a protocol like the one used byMagescas and Prablanc (2006). For the second group(SacEnd), we extinguished the target after completion ofthe orienting eye movement, reasoning that acquisition of thetarget by the eyes would render its subsequent displacement(when re-illuminated at hand touchdown) more noticeable. Forthe third group (ReachEnd), we extinguished the target coinci-dent with hand touchdown, immediately re-illuminating itat its shifted location. For the fourth group (Informed), weapplied the same stimulus protocol as for the SacStartgroup but also explicitly told participants about the pres-ence and size of the target perturbations that they would beencountering. We hypothesized that reach adaptation wouldbe most robust (i.e., show the strongest aftereVects) in theWrst group, as this is the group for which the target shouldappear most stable during the acquisition phase (Magescasand Prablanc 2006).

Methods

In the present section, we describe three experiments.We have divided the report this way to indicate that theSacStart and SacEnd groups were collected Wrst (Experiment1) and that a preliminary analysis of these data was carriedout prior to designing the protocol for the ReachEnd group(Experiment 2). The protocol for the Informed group(Experiment 3) was designed after collection and analysisof the data from Experiments 1 and 2. However, the appara-tus and procedures were identical (but for the critical exper-imental manipulations) across all groups.

Participants

For all experiments, participants were right-handed, hadnormal or corrected-to-normal vision, and were naïve to the

1 However, also see Bekkering et al. (1995). These authors show reachadaptation when pointing to an unseen target jump that occurs online,but they do not control for the eVects of saccadic adaptation and theyargue, in fact, that their results indicate transfer of saccadic adaptationto the hand. The Magescas et al. (2009) study, which deliberately pre-vented saccadic adaptation, can, therefore, be viewed as a purer test ofreach adaptation to an unseen online target perturbation.

123

Exp Brain Res (2010) 203:367–380 369

aims of the study. Participants were recruited from theUniversity of British Columbia community, and their agesranged between 18 and 36. All participants providedinformed consent prior to participating. The study met theguidelines of the university’s research ethics board and wasconducted in accordance with the guidelines set by theDeclaration of Helsinki.

Experiment 1

Twenty-two participants completed this experiment. Thedata for three participants could not be used due to errors incollection. Two participants were excluded for failure tofollow instructions. Of the remaining 17 participants, eightparticipated in the SacStart condition (3 women, 5 men)and 9 participated in the SacEnd condition (5 women, 4men).

Experiment 2

Eight participants, none of whom were involved in Experi-ment 1, completed this experiment. The results for one par-ticipant were not analyzed due to a program error thatslightly altered stimulus timing relative to the other partici-pants. The remaining seven participants (3 women, 4 men)completed the same condition (ReachEnd).

Experiment 3

Eight participants (5 women, 3 men), none of whomwere involved in Experiments 1 or 2, completed thisexperiment.

Apparatus

The apparatus was the same for all experiments. An arrayof red light-emitting diodes (LEDs) was mounted above ahalf-silvered mirror, which in turn was mounted an equaldistance above the reaching surface (Fig. 1). The opticalgeometry of the apparatus created the impression that thetargets, reXected by the mirror, were in the same plane asthe reaching hand. The half-silvered mirror allowed us toprovide or remove vision of the hand by illuminating orextinguishing, respectively, a white light below the mirror.The reaching surface was a featureless white surface but fora circular home position (5 mm diameter) located at thebottom left corner. The surface and the home position wereonly visible when the light below the mirror was turned on.

An eye Wxation point (5 mm diameter) was presented13 cm above the home position at the start of each trial(Fig. 1). The Wxation point was projected from the LEDarray and could be extinguished simultaneously with theonset of a peripheral target. Target stimuli (5 mm diameter)were also presented by illuminating LEDs in the array. Tar-gets could be presented at Wve locations, each separated byapproximately 1.25 cm. The Wrst target position waslocated 22.5 cm to the right of Wxation; the second targetposition was located 23.75 cm to the right of Wxation; andso on. Participants’ eyes were located approximately 48 cmfrom the reaching surface. Accordingly, the distancebetween adjacent targets was approximately 1.5° of visualangle.

Aiming movements were made with a handheld styluswith an infrared-emitting diode attached near the tip. Thisallowed us to track the position of the stylus with OPTOTRAK

Fig. 1 a Schematic of the apparatus. LEDs at the top of the apparatusare reXected by the half-silvered mirror. The reXected stimuli appear tobe in the same plane as the hand. Vision of the hand is manipulated byilluminating/extinguishing white LEDs mounted below the mirror.b Schematic of the stimulus display. Participants would begin each

trial with the stylus at the home position and their eyes on Wxation. Atarget would appear at one of the possible target locations, simulta-neous with the oVset of eye Wxation. The dotted lines represent reachtrajectories but do not reXect actual data

123

370 Exp Brain Res (2010) 203:367–380

(Northern Digital), which sampled at a rate of 500 Hz. Thestylus was also equipped with a microswitch tip thatallowed us to track movement start and end in real time.Electrooculography (EOG) was used to monitor horizontalsaccades. Disposable Ag–AgCl surface electrodes wereplaced at the outer canthi of the eyes with a reference elec-trode placed on the forehead. EOG signals were ampliWed(5–10 K) and band-pass Wltered (0.1–30 Hz) using an ACpreampliWer (Grass Instruments P511) and sampled at arate of 500 Hz. The EOG signal was passed through an ana-log circuit that enabled online triggering of stimulus events.For each participant, we manually set a voltage thresholdsuch that the trigger would occur within the Wrst 3rd of thesaccade.

Procedure

Experiment 1 (SacStart and SacEnd groups)

Participants began each trial with the stylus at the homeposition and their eyes at the Wxation point. When theexperimenter triggered the start of a trial, a warning beepsounded. After a variable fore-period (ranging between 1and 2 s), the Wxation point would disappear and a targetwould simultaneously appear in the visual periphery. Theparticipant was instructed to look and point, smoothlyand accurately, to the target. The light below the mirrorwas on at the start of every trial, such that the participantcould view the home position and their hand. Each

participant completed ten practice trials prior to beginningthe experiment.

The experiment was divided into four phases: a pretestphase, an acquisition phase, and two post-test phases. Thepretest phase consisted of 50 trials in which a target wouldrandomly appear at one of the Wve possible locations. Whenthe eyes began to move, the target and the light below themirror were both extinguished and remained oV until theend of the trial (i.e., participants had vision of their handand the target at the start of the trial only) (Fig. 2). After thetrial was over and the participant had begun to return to thehome position, the light below the mirror was re-illumi-nated (750 ms after lifting the stylus from its endpoint posi-tion) so that the participant could place the stylus at thehome position. After completing the pretest, the participantwas given a break as well as some additional instructions.Participants were told that the next trials would includevisual feedback, such that they would see their hand and thetarget at the end of the movement. They were instructed toapproach each trial with the intent to minimize any errorbetween their hand and target that they had seen at the endof the previous trial. These instructions were the same forboth groups of participants.

The only manipulation that diVered between theSacStart and SacEnd groups occurred in the acquisitionphase of the experiment. For both groups, the light belowthe mirror was extinguished at the onset of the orientingsaccade (speciWcally, when the EOG voltage surpassed thetrigger threshold). In the SacStart group, the target was

Fig. 2 a Sequence of stimulus events during the pre-test and post-test 1 phases of all experi-ments b Sequence of events dur-ing the acquisition and post-test2 phases of all experiments. Inset: Target behavior during the acquisition phases of all experi-ments. Note that the initial target location is the same for all acqui-sition trials, while its second (feedback) location gradually moves to the right

123

Exp Brain Res (2010) 203:367–380 371

extinguished at this point as well. In the SacEnd group,however, the target remained illuminated for 90 ms afterthe saccade trigger, allowing the eyes to acquire the target.The target was then extinguished. For both groups, the tar-get and the hand were re-illuminated for 1 s when the styluscontacted the surface at the end of the reach. The acquisi-tion phase consisted of 50 trials, and the initial (pre-reach)target location was the same for all of these trials (22.5 cmto the right of Wxation). During the Wrst ten trials, the targetreappeared at its initial location when the stylus landed(i.e., the target did not move). For the subsequent ten trials,the target reappeared 1.25 cm to the right of its initial loca-tion when the stylus landed. Every ten trials, the size of thedisplacement increased by 1.25 cm, such that for the Wnal10 trials of acquisition, the target was reappearing 5 cm tothe right of its initial location. This protocol, except for theSacEnd manipulation, is very similar to that of Magescasand Prablanc (2006).

The Wrst post-test phase was identical to the pretestphase.

The second post-test phase consisted of 20 trials thatwere identical to the Wrst ten trials of the acquisition phase.In other words, the hand and the target were visible at theend of the reach, but the target was never displaced. Thesetrials served to washout any residual adaptation that mighthave been present after the 50 trials of the Wrst post-testphase.

After completion of the experiment, participants wereasked about the target displacement in the acquisitionphase: whether they noticed it and, if so, how often itoccurred and where the target moved to.

Experiment 2 (ReachEnd group)

The procedure for this experiment was identical to Experi-ment 1 but for two diVerences: (1) in the acquisition phase,the target remained illuminated at its initial location untilthe stylus landed (at which point the light under the mirrorwas turned on and the target re-appeared at its shifted loca-tion, as in the Wrst experiment) (Fig. 2) and (2) after the Wrst11 trials of acquisition (i.e., after the participant had experi-enced 10 stationary feedback trials and the Wrst of the 40displacement trials), participants were asked whether theynoticed anything happen to the target on the last trial. Theywere then instructed to ensure that on subsequent trials theyattempt to aim such that the stylus would land as close tothe second target location as possible.

Experiment 3 (Informed group)

This group was given an identical stimulus protocol to theSacStart group of Experiment 1. In other words, during theacquisition phase, the target was extinguished at the onset

of the orienting saccade and then reappeared at hand touch-down. The critical diVerence between the Informed groupand the SacStart group was that the Informed group wastold about the presence of the target perturbation. For theWrst ten trials of the acquisition phase (when the target reap-peared at the same location from which it disappeared), par-ticipants were instructed simply to use the visual feedbackto minimize reach error on subsequent trials. After experi-encing the Wrst trial on which a perturbation occurred, par-ticipants were asked if they noticed anything and then toldthat on that trial the target had actually moved to the rightby 1.25 cm and that this would continue to occur for thenext several trials. They were also told to continue withattempting to minimize error between their hand and theWnal location of the target (i.e., to go to the target’s secondlocation). After the following ten trials, the last of whichincluded a 2.5-cm perturbation, participants were againasked if they noticed anything and then told of the increasein the size of the rightward jump. This process of informingthe participant about the increased jump magnitude every10 trials continued until the end of acquisition. After com-pleting the Wrst trial of the post-test, participants wereinformed that there would no longer be any visual feedbackat the end of their movement (like the trial they just com-pleted) and to simply aim to the target that appeared at thestart of each trial for the remainder of the trials.

Unlike Experiments 1 and 2, there was no post-test 2 forExperiment 3.

Data reduction and analysis

Because our manipulation systematically induced targeterror in the lateral dimension, we focused our analysis onthis dimension of the reaching movements. Prior to anyanalysis, movement endpoints were normalized withrespect to the actual target location (normalized lateralendpoint = raw lateral endpoint–lateral position of targetfor that trial).

We were interested in the robustness of adaptation ineach of the four groups, so our analysis focuses on the rateof de-adaptation in the Wrst post-test phase for each group.SpeciWcally, we examine the number of post-test trialsrequired for each group to reach baseline (0 mm targeterror) during the post-test phase. We do not use the pretestvalues as a baseline; as will become evident in the resultssection, many participants exhibited a large undershoot biasduring the pretest, presumably because the absence ofvisual feedback during this phase prevented them from cal-ibrating to the reach environment. This large undershootpromptly vanished when visual feedback was introducedduring the acquisition phase (Fig. 3), prior to the introduc-tion of the target perturbation. Because we were interestedin measuring the inXuence of the target perturbation (and

123

372 Exp Brain Res (2010) 203:367–380

only the target perturbation), we do not use the pretest as abaseline, as this would introduce the additional, distortingeVects of the un-calibrated movement error.

Statistical analysis

We carry out two sets of analyses on the Post-test 1 phaseto assess de-adaptation across the four groups. For our Wrstanalysis, which compares the four groups to each other, wecollapsed across every Wve trials of the post-test (i.e., trials2–6 of the post-test, trials 7–11 of the post-test, and so on)and submitted these means, along with the Wrst trial of thepost-test, to a 4 (group) £ 10 (stage of Post-test 1) ANOVAwith repeated measures on the second factor. For our sec-ond analysis, which examines the rate at which each groupreaches baseline in the post-test, we took the same Wve-trialmeans used in the Wrst analysis and compared each mean to0 with a one-sample t-test. The second analysis involvesnine t-tests for each group, for a total of 36 t-tests. Becausethis was primarily a way to objectively determine when agroup had ceased to signiWcantly overshoot the target (i.e.,a search for an absence of a diVerence), we were not overlyconcerned about the inXated risk of type-I error and there-fore used an alpha of .05 for each comparison (Aivar et al.2008).

Violations to sphericity on any repeated measures factorwere addressed with a Huynh–Feldt correction to degreesof freedom.

Comparisons of endpoint variability between groupswere tested with the Bartlett test for homogeneity ofvariance.

Finally, we note that, although we have described Exper-iments 1, 2, and 3 separately in the methods section, thedata from all experiments are considered together in the fol-lowing results section.

Results

Eye, hand, and target timings

The average start and end times for the eye and hand aredisplayed in Table 1. Additionally, the table shows the tim-ing of target oVset with respect to the eye. These resultsindicate that, on average, the eyes led the hand by 61 ms,the target disappeared 30 ms after the onset of the saccadein the SacStart and Informed groups, and the target disap-peared 28 ms after completion of the saccade in the SacEndgroup.

Motor adaptation to the target perturbation

Figure 3 shows the adaptation and de-adaptation to the tar-get displacement across the 4 groups: SacStart, SacEnd,ReachEnd, and Informed. Visual inspection of Fig. 3 sug-gests that, while all groups modiWed their reach endpointsduring the acquisition phase in response to the increasingrightward target displacement, the ReachEnd group andthe Informed group de-adapted (Post-test 1) faster thanthe other groups. A statistically signiWcant interactionbetween the stage of Post-test 1 factor and the group factorconWrms that not all groups de-adapted at the same rate,

Fig. 3 Adaptation proWle for all 4 groups. All data points, except theWrst of the post-test phase, are data that have been collapsed across 5consecutive trials. The Wrst data point in the post-test reXects only theWrst trial of the post-test. The numbers indicate the onset of visual feed-back (1), the onset of the target perturbation (2), and the oVset of both

the perturbation and visual feedback (3). Endpoint error (on the y-axis)is with respect to the initial target location for each trial. Note that theinitial (pre-saccadic) target location was the same for all trials of theacquisition phase

123

Exp Brain Res (2010) 203:367–380 373

F(12.2, 113.9) = 2.34, P = .0098. Comparisons between themeans at each stage of the post-test and a baseline of 0-mmerror show that the ReachEnd group was fastest to reachbaseline, followed by the Informed group, then the SacStartand SacEnd groups (Fig. 4). The ReachEnd group reachedbaseline within six trials and the Informed group within 11.The SacStart and SacEnd groups, on the other hand,remained signiWcantly diVerent from baseline until, at least,post-test trials 31 and 46, respectively.

Interestingly, the ReachEnd group appeared to develop asigniWcant target undershoot during the post-test, the possi-ble implications of which are discussed later.

Awareness of the target perturbation

In the SacStart group, seven of the eight participantsreported no awareness of target displacement. In theSacEnd group, Wve of the nine participants reported no

Table 1 Timing of various events (in ms) for each stage of each group

Saccade onset

Trigger Reach onset

Saccade end

T1 oV Reach end

Reach MT

SacStart

Pretest 241.66 270.65 288.78 337.86 270.65 661.33 372.54

Acquisition 231.37 261.53 290.73 326.99 261.53 677.27 386.54

Post-test1 233.24 264.08 273.83 339.45 264.08 684.52 410.69

SacEnd

Pretest 217.83 247.51 280.44 309.36 337.51 687.39 406.95

Acquisition 202.33 230.26 275.24 291.75 320.26 658.41 383.17

Post-test1 211.47 239.68 261.07 309.37 329.68 678.77 417.70

ReachEnd

Pretest 207.85 236.67 263.84 294.73 670.63 670.63 406.79

Acquisition 198.32 226.66 266.64 281.81 688.09 688.09 421.45

Post-test1 200.48 229.86 254.19 292.95 673.17 673.17 418.98

Informed

Pretest 232.03 265.29 309.98 315.63 265.29 839.92 529.94

Acquisition 237.61 269.95 320.04 319.51 269.95 839.61 519.57

Post-test1 228.01 262.33 290.45 320.52 262.33 829.06 538.61Values represent group means

Fig. 4 De-adaptation during Post-test 1. Group means are plotted for each 5 trials of the post-test, starting with the sec-ond trial of the post-test. Error bars represent between-subject SD. The statistical signiWcance of one-sample t-tests comparing each mean to 0 mm is indicated with asterisks: ** P < .01; * P < .05

123

374 Exp Brain Res (2010) 203:367–380

awareness of target displacement. The two groups did notdiVer signiWcantly from each other on this measure,�2 = 2.08, P = .149. This was surprising, as we hadexpected the delayed oVset of the initial target location inthe SacEnd group to dramatically increase awareness of theperturbation. Even for those participants who did reportawareness of the displacement, awareness appeared to beincomplete. When asked, these participants reported, onaverage, that the target changed location on about 45% ofthe acquisition trials (The actual proportion was 80%). Itshould be noted that the verbal report was a very rough esti-mate of awareness, because questioning about the acquisi-tion phase came at the end of testing (after 70 interveningpost-test trials) and also because the verbal report onlymeasured what participants thought they detected duringtesting; it was not a measure of what participants are capa-ble of detecting when they actively search for a perturba-tion.

In the ReachEnd group, participants were asked whetherthey noticed anything happen after encountering the Wrsttrial in which a target displacement occurred. Five of sevenparticipants reported that the target jumped right. The othertwo participants, surprisingly, did not notice the Wrst jump,but both noticed it on the next trial. All participants were,therefore, aware of the jump to some extent. When ques-tioned at the end of the experiment, Wve of the participantswere able to correctly report that the size of the jump hadincreased during the acquisition period. The other two par-ticipants thought that jump size was consistent throughoutacquisition, and both estimated it to be ‘about one inch’.

In the Informed group, participants were asked whetherthey noticed anything happen after the Wrst trial in which atarget displacement occurred and then again each time themagnitude of the displacement increased. No participantsnoticed the Wrst target displacement or the increase in dis-placement size from 1.25 to 2.5 cm. When the displacementincreased to 3.5 and 5 cm, several participants successfullyindicated that it had moved farther to the right, but it isunclear how much of this report reXects perception of anincreased displacement magnitude vs. recognition of thepattern of target movement reported by the experimenter.Informal questioning of the participants after completion ofthe experiment suggested that most participants felt likethey could see the largest (5 cm) displacement but had diY-culty perceiving the displacement at smaller magnitudes.

Endpoint variability in the SacStart and SacEnd groups

Greater inter-subject variability in the post-test for onegroup relative to another may reXect less robust adaptation.Indeed, the apparently larger standard deviation error barsfor the SacEnd group relative to the SacStart group (Fig. 4)are suggestive of less robust adaptation in the SacEnd

group. However, statistical comparison of the variability ofthe SacStart post-test mean (SD: 12.3) to the variability ofthe SacEnd post-test mean (SD: 21.7) did not reveal a sig-niWcant diVerence, �2(1) = 2.09, P = .148.

Pretest diVerences among groups

Although we do not use the pretest as a baseline for mea-suring adaptation (due to its inclusion of un-calibratedmovement error that might distort measurement of targetperturbation eVects), there remains the possibility thatdiVerences among groups in the post-test (which we haveattributed to experimental manipulations) might be inXu-enced by initial diVerences among groups in the pretest.Comparison of the four groups’ pretest means (Fig. 5) sug-gests that the Informed group’s pretest mean may be sig-niWcantly lower than the SacEnd pretest mean (though notthe other groups’ pretest means).2

In light of this, we examined the performance of theInformed group in greater detail. We asked whether a par-ticipant’s mean endpoint in the pretest was related to theirrate of de-adaptation in the post-test. A signiWcant negativecorrelation between pretest mean and the amount of decayin the early post-test (i.e., lower pretest value associatedwith larger decay value) might suggest that diVerences inrates of de-adaptation between the Informed group andother groups (especially the SacEnd group) should not beattributed to the experimental manipulations alone. How-ever, there was no signiWcant correlation between pretest

2 Violation of the assumption of homogeneity of variance (HOV)among groups on the pretest (Bartlett, P = .0075), along with unequalns among the groups, dictates that a non-parametric test should be ap-plied. When a Kruskal–Wallis test is applied to the groups, the result isnot signiWcant, H(3, N = 32) = 6.49, P = .09, However, if one over-looks violation of HOV and applies a one-way ANOVA instead, a sig-niWcant diVerence does emerge, F(3,28) = 2.96, P = .049. Subsequentpost-hoc testing with Newman-Keuls reveals that the SacEnd andInformed groups do diVer signiWcantly in the pretest, P = .04. All otherpairwise contrasts are not signiWcant (P > .10).

Fig. 5 Mean lateral endpoint error in the pretest phase for each group.Error bars represent between-subject SD

123

Exp Brain Res (2010) 203:367–380 375

mean and endpoint decay from post-test trial 1 to post-testtrials 2–6, r = ¡.03, P > .90, or between pretest mean andendpoint decay from post-test trial 1 to post-test trials 7–11,r = ¡.05, P > .90. These near-zero correlations suggest thatperformance on the pretest in the Informed group is notrelated to their rate of de-adaptation in the early post-test; itshould, therefore, be safe to make comparisons betweenthis group and the other groups despite its potentially lowerpretest mean.

Discussion

Our goal was to examine the inXuence of perceived targetstability on reach adaptation, and we hypothesized thatadaptation would decrease with decreased stability. Weattempted to manipulate the perceived stability of the targetby altering the time at which it disappeared relative to theorienting eye movement, reasoning that the shorter theinterval between the oVset of the target at its initial locationand the onset of the target at its displaced location, the lessstable it would appear. We hypothesized, therefore, thatparticipants exposed to the earliest oVset of the target at itsinitial location would show the strongest aftereVects. Wealso investigated the impact of informing participants aboutthe presence of the target perturbation on the robustness ofadaptation to the perturbation. In total, we compared adap-tation to target error across four groups: one in which thetarget disappeared shortly after the onset of the saccade(SacStart), one in which the target disappeared shortly afterthe end of the saccade (SacEnd), one in which the targetdisappeared at the end of the reach (ReachEnd), and one inwhich participants were informed that the target was beingperturbed during their reach (Informed).

Our results failed to show the predicted diVerence in theaftereVects of the SacStart and SacEnd groups. Both groupsexhibited a large and long-lasting rightward shift of reach-ing movements in the post-test, whereas we had predictedmore rapid de-adaptation in the SacEnd group. The Reac-hEnd and Informed groups, by contrast, exhibited fasterrates of de-adaptation in the post-test. The timing of a targetperturbation and a participant’s knowledge of a target per-turbation appear, therefore, to inXuence aftereVect robust-ness.

It is important to note that all groups received the sameterminal feedback during the acquisition phase: For allgroups, the hand was illuminated at the same time that thetarget appeared at its displaced location. The more rapidaftereVect decay in the ReachEnd group is, therefore, due tosome change in the way the terminal error was interpretedby the visuomotor system as a result of the timing of theinitial target’s oVset. This change in timing was designed tomake the perturbation more explicit, and so participant

awareness of the target perturbation in this group mayexplain the short-lived aftereVects. (Indeed, the results ofthe Informed group suggest that awareness of the perturba-tion can reduce aftereVect duration). However, it is alsopossible that reach error was processed altogether diVer-ently in the ReachEnd group because the initial targetremained illuminated throughout the duration of the reach.In the following sections, we discuss the possible inXuencesof perturbation awareness and perturbation timing on reachadaptation.

Awareness of the target perturbation

We suggest that a key factor in the decay of the motor after-eVect was participant awareness of the target perturbation.This suggestion rests on the Wnding that the Informed andthe ReachEnd groups (in which participants were aware ofthe target perturbation) exhibited faster de-adaptation ratesthan the other two groups (in which participants werelargely to partially unaware). However, limitations in ourmeasure of awareness as well as changes in the timing oftarget oVset across the groups do present challenges for thisexplanation, as we discuss below.

In Experiment 1, we had hypothesized that delaying thetarget’s disappearance in the SacEnd group would (1)increase participant awareness of the perturbation and (2)decrease adaptation to the same perturbation relative to theSacStart group. The results of this comparison, however,were inconclusive. The SacEnd group did not show a sub-stantial increase in awareness relative to the SacStart group(as measured by post-experiment verbal report) nor didthey exhibit signiWcantly diVerent behavior in the post-testphase relative to the SacStart group. The lack of eVectbetween the SacEnd and SacStart groups, therefore, maysimply reXect a failure in our manipulation to increaseinstability of the target. Without a more sensitive measureof participant awareness, however, we cannot conWrm this.

The performance of the ReachEnd group (Experiment2), on the other hand, is suggestive of a role for awarenessin the adaptation to the target perturbation. This groupexperienced a much later oVset of the target and we con-Wrmed, during the experiment, that this perturbation wasconsciously visible to the participants. Under these condi-tions, participants exhibited much faster decay of after-eVects relative to the other groups. However, because wewere simultaneously manipulating visibility of the targetperturbation and the timing of the target’s disappearancerelative to the reach, it is possible that some eVect of thedelayed oVset other than an increase in awareness wasresponsible for the rapid de-adaptation in this group.

It may be, for instance, that the longer visual persistenceof the target at its initial location in the ReachEnd conditionallowed the visuomotor system to process reach error with

123

376 Exp Brain Res (2010) 203:367–380

respect to that location rather than the post-perturbationlocation. This might have occurred online or at movementtermination. Given the absence of hand vision (recall thatthe hand was only visible when the perturbed target loca-tion was presented), any online error processing would berelying on an internal estimate of hand location for hand-to-target error processing. Such error processing is by nomeans implausible; online corrections to unseen target per-turbations in the absence of limb vision have been well doc-umented (e.g., Goodale et al. 1986; Pelisson et al. 1986;Desmurget et al. 1999; Bard et al. 1999). However, if sucherror processing was occurring, it was eVectively overrid-den by the participants during the acquisition phase. Notonly did participants terminate their movements at the per-turbed location, but they also prepared their movements tothe perturbed location during the acquisition phase.3 Inother words, any online error processing with respect to theinitial target location either did not occur or, if it did, wascountermanded by reach end; more importantly, any oZineinXuence of such error processing did not appear to inXu-ence subsequent movement preparation during the acquisi-tion phase.

The other possibility is that terminal error processingwas inXuenced by the persistence of the initial target loca-tion. The persistence of the initial location in the ReachEndgroup may have forced participants into a cognitive modeof control (e.g., ‘reach further right’) that prevented errorprocessing with respect to the post-perturbation location(although as we discuss later, we believe cognitive controlunder these circumstances to be more of an eVect than acause of failed adaptation). An interesting side eVect ofconsistently overshooting the initial target location in theReachEnd group may have been the implicit processing oferror with respect to that location rather than the Wnal targetlocation. It is possible that upon re-illumination of the handat movement end, error was processed with respect to therecently extinguished initial target location. This consistentovershoot error during acquisition might then have led to anundershoot error in the post-test, an eVect that appeared toemerge in our results for the ReachEnd group. Of course,our experiment was not designed to test this possibility, andso the occurrence of such error processing remains largelyspeculative.

An alternative way to consider the ReachEnd results iswithin the context of diVerent timescales of adaptation.Smith et al. (2006) have suggested that fast-adapting andslow-adapting processes might simultaneously be engagedduring motor adaptation, and they have shown that a modelthat considers multiple timescales closely mimics the

behavior of participants engaged in learning and relearninga force-Weld perturbation task. In their model, any learningthat occurs in the fast-adapting system decays rapidly whenthe perturbation is removed (or a counter-perturbation isintroduced), whereas the slow-adapting system retains itslearning over an extended period (up to several hundred tri-als, depending on the duration of the prior exposure).Although Smith et al. do not discuss whether the two pro-cesses can be engaged independently, one possibility is thatour ReachEnd group only engaged fast-adaptive processingand that this is why the group exhibited such rapid extinc-tion during the post-test. This would still leave the questionwhy this group only experienced fast-adaptive processing.It would be tempting to draw parallels between so-calledcognitive control and fast-adaptive processes (Indeed, ini-tial error reduction during prism exposure has generallybeen attributed to cognitive control (Redding et al. 2005;Michel et al. 2007). However, Smith et al. argue that a fast-adaptive system would probably reside in, or at least relyon, the cerebellum, suggesting that cognitive and fast-adap-tive processes are not the same thing.

There is also the possibility that two processes weresimultaneously occurring for the ReachEnd group: anexplicit form of learning about the target perturbation thatrapidly decayed in the post-test and an implicit form oferror processing with respect to the initial target locationthat led to the emergence of a negative aftereVect. Again,however, we can only speculate that such implicit process-ing was occurring, and it is not clear whether such explicitand implicit processes are directly comparable to Smithet al.’s (2006) fast- and slow-adaptive processes.

Even if we do posit implicit error processing of the ini-tial target location for the ReachEnd group, we are still lefttrying to understand why these participants failed to implic-itly learn from the perturbed target location. An explana-tion that eschews any consideration of awareness wouldhave to posit some mechanism by which, at a certain pointin the reach (sometime after saccade completion), availabletarget information (i.e., the initial target location in theReachEnd group) is used to the exclusion of target informa-tion presented at the end of the reach (i.e., the perturbedlocation in the ReachEnd group). Such a mechanism maybe operating, but we currently have no theoretical frame-work from which to judge its plausibility. A more likelyexplanation, we believe, is that visibility of a perturbationimmediately disqualiWes the perturbed location of the targetas a viable landmark for error processing, thereby openingthe possibility of processing reach error with respect to theinitial target location.

With Experiment 3 (Informed group), we tested theeVect of participant awareness of the perturbation withoutthe confounding eVects of a persisting initial target loca-tion. This group exhibited a faster rate of de-adaptation

3 Based on the Wnding that peak velocity linearly increased as the per-turbation size increased across acquisition trials (r = .86),F(1,48) = 133.62, P < .001.

123

Exp Brain Res (2010) 203:367–380 377

than the SacStart and SacEnd groups, suggesting thatawareness that a perturbation is occurring can undermineadaptation. This is a result that parallels Wndings from theprism literature, in which it has been shown that awarenessof a perturbation reduces adaptation (Jakobson andGoodale 1989; Michel et al. 2007). Unlike those studies, how-ever, we introduced an unstable target rather than directlyaltering reaVerence. It remains unclear whether our result inExperiment 3 is caused by an increased visibility of the tar-get perturbation as a result of being told that it is occurringor the result of cognitive control induced by the instruction.

When people are told that a perturbation is occurring,this might have the eVect of increasing the visibility of theperturbation in a top–down manner (somewhat analogousto a person’s ability to extract a coherent percept from anambiguous image only after being told what to look for(Coren et al. 1999)). Thus, a target displacement that mighthave otherwise gone unperceived might become percepti-ble as a result of instructions that direct one’s attention to it.Although we would not expect top–down processes to dra-matically inXuence the perceptibility of an intra-saccadictarget displacement, which is subject to strong saccadicsuppression eVects, the post-saccadic displacementemployed here may be more susceptible to detection whenit is actively sought (Deubel et al. 1996). However, partici-pants’ post-experiment reports of perceptibility in Experi-ment 3 do suggest that, even when informed about thepresence of the displacement, they had diYculty perceivingit.

In short, there are two possible awareness-related expla-nations for the reduced adaptation in the ReachEnd andInformed groups. In what we will term the ‘bottom-up’explanation, visibility of a target displacement might under-mine the motor system’s assumption of a stable environ-ment, causing it to attribute reach error to the environmentrather than an incorrect motor command.4 Alternatively, inwhat we will term the ‘top–down’ explanation, awarenessthat a perturbation is occurring might prompt a cognitive,or explicit, mode of reach control that potentially inhibitsimplicit motor adaptation. It is possible that both explana-tions apply, and they may even apply diVerently to theReachEnd and Informed groups. In the following section,we discuss previous research on explicit and implicit pro-cesses in motor adaptation, research that suggests displace-ment visibility may be a more important determinant ofmotor adaptation than cognitive control.

Implicit vs. explicit processes in motor adaptation

Previous research suggests that a person’s awareness of aperturbation impacts how they learn from it, at least when itcomes to prism adaptation. Michel et al. (2007), forinstance, showed that awareness of a prismatic displace-ment decreases the magnitude and robustness of after-eVects. They also found that awareness of the perturbationdecreased inter-limb transfer of the adaptation. In theirstudy, Michel et al. compared a group that received imme-diate onset of a consciously detectable prismatic perturba-tion of 10° (aware group) to a group that received anincremental, and undetectable, increase in prismatic pertur-bation from 2 to 10°s (unaware group). Overall, partici-pants in the aware group exhibited 40% adaptation to thedistortion compared to 56% for the unaware group. Otherresearch has also shown that when participants are toldabout the presence of a 5° prism distortion, they adapt lessthan participants who are not told about the presence of thesame distortion (Jakobson and Goodale 1989).

Such Wndings are partially consistent with our sugges-tion that awareness played a role in the rapid de-adaptationof our ReachEnd and Informed groups. We say ‘partiallyconsistent’, because there is reason to believe that a detect-able target perturbation undermines adaptation more than adetectable prism shift or cursor rotation. For instance, Died-richsen et al. (2005) (described in the Introduction) showedthat people adapt more to a 12° cursor rotation than they doto a comparable online target jump. Although Diedrichsenet al. did not measure awareness of perturbations in theirstudy, it is safe to assume that the 12° cursor rotation andthe non-saccadically suppressed target jump were easilydetected by participants. Moreover, an important feature ofthe Diedrichsen et al. (2005) study is that they employed arandomized perturbation procedure, such that participantscould not anticipate the upcoming jump/rotation direction.The eVects that they observed, therefore, would not likelybe due to cognitive (i.e., top–down) control of reaching.Any adaptation (or lack thereof) that they observed would,rather, be due to lower-level processing. Their results,therefore, are consistent with the idea that visibility of a tar-get displacement can dramatically inXuence motor adapta-tion, independent of the presence/absence of cognitivecontrol. We do not mean to imply, however, that cognitivecontrol of the reach was absent in our ReachEnd andInformed groups; it was very likely required, at least in theReachEnd group, for successfully hitting the perturbed tar-get during the acquisition phase. Rather, we are suggestingthat cognitive control likely emerged as an eVect of the fail-ure to learn implicitly, instead of being the cause of the fail-ure.

Work by Mazzoni and Krakauer (2006) provides furtherevidence that cognitive control per se does not prevent

4 This ‘bottom-up’ explanation does not preclude the possibility thattop–down factors (such as instruction in the Informed group) may ini-tially induce visibility of a target displacement. If visibility were beinginduced in this way, we would still consider the bottom-up explanationto account for the reduced adaptation.

123

378 Exp Brain Res (2010) 203:367–380

motor adaptation. In their study, participants who weregiven an explicit strategy by which they could successfullyguide a rotated cursor to a target still exhibited robustimplicit adaptation to the rotation. In other words, partici-pants who were cognitively engaged in aiming to one target(reach target) so that the rotated cursor would hit an adja-cent target (cursor target) gradually worsened at hitting thecursor target, eVectively adapting to the cursor error aroundthe intended reach target. These participants also exhibitedlong-lasting aftereVects, suggesting that implicit motorlearning is not undermined by cognitive control of thereach. Rather, implicit processes override explicit controlunder these conditions.

The foregoing evidence suggests that the short-livedaftereVects in our ReachEnd group, and possibly theInformed group, are the result of low-level registration ofan unstable reaching environment, rather than the result ofthe cognitive control that may have been occurring simulta-neously.

Relevance of reach error

Why would the motor system fail to adapt to a visible targetdisplacement? The simple answer is that learning from sucherror would be counterproductive. If the error can be attrib-uted to the environment rather than the motor system, mod-ifying the motor system should normally serve littlepurpose. Wei and Kording (2009) have recently shown thatthe amount that reach error is weighted during motor learn-ing depends on the size of the error. Larger errors receiveless weighting, apparently because the likelihood that theyare due to planning error (as opposed to the environment) islower than for smaller errors. In other words, large errorsseem to be less relevant for learning than small errors.

Unlike Wei and Kording’s (2009) study, our studymanipulated visibility of a target perturbation, rather thanthe size of the hand-to-target error. Accordingly, a poten-tially important diVerence between our study and that ofWei and Kording is the basis upon which error relevance isbeing assessed by the participant. In Wei and Kording’sstudy, relevance was based on a post-reach comparison oftarget and reach endpoint. In our study, by contrast, visibil-ity of a target perturbation may have reduced relevanceprior to any comparison of the reach endpoint to the target.In fact, hand-to-target comparisons in our study would havebeen roughly equivalent across the four groups. Further-more, the incremental nature of the target error in our study(increments of 1.25 cm) was such that the size of the erroron any given trial would have been well within a range thatwould allow it to be self-attributed. Thus, while we con-sider our Wndings to be consistent with the error-relevanceperspective of Wei and Kording, the visibility of a targetperturbation in the ReachEnd group may have reduced the

likelihood that reach error was due to planning error to avery low (perhaps near-zero) level. If, however, visibility ofa target perturbation could be manipulated in a graded fash-ion (analogous to the graded spatial error used by Wei andKording), we would predict an inverse relationship betweena person’s conWdence that a perturbation had occurred andthe amount of motor adaptation resulting from that trial.

The mechanisms of motor adaptation to target error

A prominent theory of motor adaptation suggests that adap-tation relies on a comparison of predicted sensory feedbackof the movement to actual sensory feedback (e.g. Wolpertet al. 2001; Tseng et al. 2007). This theory proposes that acopy of the outgoing motor command is fed through a for-ward model (an internal representation of an eVectorsystem), which generates a prediction of the sensory conse-quences of the motor command. Any diVerence betweenthis prediction and the actual outcome can then be used as atraining signal for an internal model (Miall and Wolpert1996; Ramnani 2006).

However, in the case of a target perturbation (Magescasand Prablanc 2006), there is seemingly no conXict betweenpredicted and actual feedback of the hand because vision ofthe hand is not distorted (unlike what occurs with prisms ora cursor rotation). On this basis, Magescas et al. (2009)have challenged the notion that motor adaptation relies on acomparison of prediction and feedback.

Another way to view the target, however, is as an envi-ronmental reference point, especially when it is the primaryvisual feature of the reach environment (as it is in Magescasand Prablanc (2006) and the present study). If we reachtoward that point (while simultaneously predicting that ourreach will land somewhere near that point), only to Wnd thatour hand did not land as near to the point as we had pre-dicted, we would experience conXict between predicted andactual feedback. Although egocentrically there would be noconXict between predicted and actual feedback, allocentri-cally there would be. In other words, an imperceptible tar-get displacement could plausibly introduce disagreementbetween predicted and viewed feedback of the hand. If weconsider the results of the present study in this context, thevisible target perturbation that occurred for the ReachEndgroup would undermine the visual system’s assumption ofenvironmental stability and, accordingly, prevent the visuo-motor system from registering a mismatch between pre-dicted and viewed feedback. This would, in turn, preventany internal model updating.

Neural correlates of reach adaptation

Both the posterior parietal cortex (PPC) and the cerebellumhave been implicated in the adaptation of reaching move-

123

Exp Brain Res (2010) 203:367–380 379

ments to laterally displacing prisms, and it has been sug-gested that diVerent adaptive processes are served by thediVerent areas. The strategic component of adaptationmight rely more on the PPC, while the spatial realignmentcomponent might rely more on the cerebellum (Pisella et al.2004). The importance of the cerebellum in spatial realign-ment is supported by work showing that cerebellar lesionscan produce reduced aftereVects in both humans (Weineret al. 1983) and macaques (Baizer et al. 1999). Tseng et al.(2007) have also shown that cerebellar patients exhibit notonly reduced aftereVects but also impaired error correctionduring the exposure phase. Bilateral lesions of the PPC, onthe other hand, do not appear to impair the development ofaftereVects (Pisella et al. 2004). Neuroimaging studies,meanwhile, have implicated the PPC alone (Clower et al.1996) or both the PPC and cerebellum (Danckert et al.2008) in the early stages of prism exposure, while a recentstudy by Luaute et al. (2009) that examined extended prismexposure has implicated the PPC in the early stages ofadaptation and the cerebellum in the later stages of adapta-tion. Taken together, the foregoing evidence suggests thatthe cerebellum is critical for post-exposure aftereVects butmay also be involved in error reduction during the exposurephase, while the PPC is mostly involved in the early, possi-bly strategic, component of error reduction.

Although our study did not employ prisms, it is likelythat some of the processes involved in prism adaptation arealso involved in target-error adaptation. Thus, it is reason-able to speculate that our SacStart and SacEnd groups,which both exhibited robust aftereVects, developed cerebel-lum-based internal model adjustment over the course of theacquisition phase. The ReachEnd and Informed groups,however, which were both aware of the perturbation andwhich exhibited rapid extinction in the post-test, probablyrelied on strategic, PPC-based processes throughout theacquisition phase. The absence of aftereVects in these twogroups may indicate less involvement of cerebellum-basedlearning.

The study by Diedrichsen et al. (2005) may also beinformative in this context, as it examined the neural corre-lates of target errors (target jump) vs. execution errors (cur-sor rotation) and showed greater activation in the PPC fortarget errors than for execution errors. The target errors inthat study were created with visible target jumps, makingthem comparable to the explicit target errors experiencedby our ReachEnd and (possibly) Informed groups. Thislends further support to the speculation that PPC-based pro-cesses dominated in these two groups.

Conclusion

We manipulated the timing of target oVset and observed theadaptation to post-reach error as a function of this timing.

We also informed some participants about the presence of atarget perturbation. Our results show that (1) participantswho experienced the latest oVset (reach end) exhibited theleast amount of adaptation and (2) participants who wereinformed about the target perturbation also exhibitedreduced adaptation. We suggest that the later the oVset ofthe target, the more unstable the target appeared, causingthe motor system to attribute error to the environmentrather than itself. In future experiments, we will examinewhether it is top–down control or bottom-up registration ofmore than one target location that is directly responsible forthe rapid de-adaptation we observed in these groups.

Acknowledgments This research was supported by a NaturalSciences and Engineering Research Council of Canada (NSERC)Discovery Grant awarded to R. Chua. We would also like to thank twoanonymous reviewers for their helpful comments on the manuscript.

References

Aivar MP, Brenner E, Smeets JBJ (2008) Avoiding moving obstacles.Exp Brain Res 190:251–264

Baizer JS, Kralj-Hans I, Glickstein M (1999) Cerebellar lesions andprism adaptation in macaque monkeys. J Neurophysiol 81:1960–1965

Bard C, Turrell Y, Fleury M, Teasdale N, Lamarre Y, Martin O (1999)DeaVerentation and pointing with visual double-step perturba-tions. Exp Brain Res 125:410–416

Bekkering H, Abrams RA, Pratt J (1995) Transfer of saccadic adapta-tion to the manual motor system. Hum Mov Sci 14:155–164

Clower DM, HoVman JM, Votaw JR, Faber TL, Woods RP, AlexanderGE (1996) Role of posterior parietal cortex in the recalibration ofvisually guided reaching. Nature 383:618–621

Coren S, Ward LM, Enns JT (1999) Sensation and perception, 5th edn.Harcourt Brace, Fort Worth, pp 310–312

Danckert J, Ferber S, Goodale MA (2008) Direct eVects of prismaticlenses on visuomotor control: an event-related functional MRIstudy. Eur J Neurosci 28:1696–1704

Desmurget M, Epstein CM, Turner RS, Prablanc C, Alexander GE,Grafton ST (1999) Role of the posterior parietal cortex in updatingreaching movements to a visual target. Nat Neurosci 2:563–567

Deubel H, Schneider WX, Bridgeman B (1996) Postsaccadic targetblanking prevents saccadic suppression of image displacement.Vis Res 36:985–996

Diedrichsen J, Hashambhoy Y, Rane T, Shadmehr R (2005) Neuralcorrelates of reach errors. J Neurosci 25:9919–9931

Goodale MA, Pelisson D, Prablanc C (1986) Large adjustments invisually guided reaching do not depend on vision of the hand orperception of target displacement. Nature 320:748–750

Jakobson LS, Goodale MA (1989) Trajectories of reaches to prismati-cally-displaced targets: evidence for “automatic” visuomotorrecalibration. Exp Brain Res 78:575–587

Luaute J, Schwartz S, Rossetti Y, Spiridon M, Rode G, Boisson D,Vuilleumier P (2009) Dynamic changes in brain activity duringprism adaptation. J Neurosci 29:169–178

Magescas F, Prablanc C (2006) Automatic drive of limb motor plasticity.J Cogn Neurosci 18:75–83

Magescas F, Urquizar C, Prablanc C (2009) Two modes of error pro-cessing in reaching. Exp Brain Res 193:337–350

Mazzoni P, Krakauer JW (2006) An implicit plan overrides an explicitstrategy during visuomotor adaptation. J Neurosci 26:3642–3645

123

380 Exp Brain Res (2010) 203:367–380

Miall RC, Wolpert DM (1996) Forward models for physiological mo-tor control. Neural Netw 9:1265–1279

Michel C, Pisella L, Prablanc C, Rode G, Rossetti Y (2007) Enhancingvisuomotor adaptation by reducing error signals: single-step(aware) versus multiple-step (unaware) exposure to wedgeprisms. J Cogn Neurosci 19:341–350

Pelisson D, Prablanc C, Goodale MA, Jeannerod M (1986) Visual con-trol of reaching movements without vision of the limb II. Evi-dence of fast unconscious processes correcting the trajectory ofthe hand to the Wnal position of a double-step stimulus. Exp BrainRes 62:303–311

Pisella L, Michel C, Grea H, Tilikete C, Vighetto A, Rossetti Y (2004)Preserved prism adaptation in bilateral optic ataxia: strategic ver-sus adaptive reaction to prisms. Exp Brain Res 156:399–408

Ramnani N (2006) The primate cortico-cerebellar system: anatomyand function. Nat Rev Neurosci 7:511–522

Redding GM, Rossetti Y, Wallace B (2005) Applications of prismadaptation: a tutorial in theory and method. Neurosci Biobehav R29:431–444

Shadmehr R, Mussa-Ivaldi FA (1994) Adaptive representation ofdynamics during learning of a motor task. J Neurosci 14:3208–3224

Smith MA, Gazizadeh A, Shadmehr R (2006) Interacting adaptive pro-cesses with diVerent timescales underlie short-term motor learn-ing. PLOS Biol 4:1035–1043

Tseng Y, Diedrichsen J, Krakauer JW, Shadmehr R, Bastian AJ (2007)Sensory prediction errors drive cerebellum-dependent adaptationof reaching. J Neurophysiol 98:54–62

Wei K, Kording K (2009) Relevance of error: what drives motor adap-tation? J Neurophysiol 101:655–664

Weiner MJ, Hallett M, Funkenstein HH (1983) Adaptation to lateraldisplacement of vision in patients with lesions of the centralnervous system. Neurology 33:766–772

Welch RB (1986) Adaptation of space perception. In: BoV KR,Kaufman L, Thomas JP (eds) Handbook of perception and humanperformance. Wiley, New York, pp 24.1–24.45

Wolpert DM, Ghahramani Z, Flanagan JR (2001) Perspectives andproblems in motor learning. Trends Cogn Sci 5:487–494

123