Available online at www.sciencedirect.com

www.elsevier.com/locate/brainres

b r a i n r e s e a r c h 1 5 2 8 ( 2 0 1 3 ) 4 2 – 4 8

0006-8993/$ - see frohttp://dx.doi.org/10.

Abbreviations: SVnCorresponding aE-mail addresses:

Research Report

Plasticity in the adult oculomotor system: Offlineconsolidation phase gains in saccadesequence learning

Noya Meitala,b,n, Sebastian Peter Korinthc,d, Avi Karnia,e

aThe E.J. Safra Brain Research Center for the Study of Learning Disabilities, University of Haifa, Mt. Carmel, 31905, IsraelbDepartment of Neurobiology, Weizmann Institute of Science, Rehovot 76100, IsraelcDepartment of Neurocognitive Psychology, Institute of Psychology, Goethe University, Frankfurt, GermanydCenter for Individual Development and Adaptive Education of Children at Risk (IDeA), Frankfurt 60486, GermanyeThe Sagol Department of Neurobiology & Ethology, University of Haifa, Mt. Carmel, 31905, Israel

a r t i c l e i n f o

Article history:

Accepted 8 July 2013

When do adults gain in learning an oculomotor sequence? Here we show that oculomotor

training can result not only in performance gains within the training session, but also induce

Available online 15 July 2013

Keywords:

Oculomotor learning

Consolidation

Saccade

Fixation

Learning phase.

nt matter & 2013 Elsevie1016/j.brainres.2013.07.01

S, successive volitionaluthor at: Department ofNoya.Meital@weizmann

a b s t r a c t

robust offline gains in both speed and accuracy. Participants were trained and tested over two

consecutive days to perform a sequence of successive saccades. Saccades were directed to four

target letters, presented simultaneously at fixed positions. A two alternative-forced choice

question, after each trial, ensured that all targets were perceived. Eye tracking measures were

tested at the beginning and end of the training session as well as at 24 h post-training. Practice

resulted in within-session gains in accuracy and a reduction of target fixation duration

(although total trial duration remained unchanged). In addition, the total average path length

traveled by the eye increased, reflecting a decrease in undershoot saccades. At 24 h post-

training, however, additional gains were expressed in both speed and accuracy of performance;

the total trial duration as well as the fixation-position-offsets and the number of corrective

saccades decreased. The expression of delayed gains indicates offline skill consolidation

processes in the eye-movement control system. Our results show that the optimization of

some aspect, specifically saccade speed parameters, of oculomotor sequence performance

evolves mainly offline, during the post-training consolidation phase, a pattern suggestive of

learning in an expert system.

& 2013 Elsevier B.V. All rights reserved.

r B.V. All rights reserved.3

saccades; AOI, area of interestNeurobiology, Weizmann Institute of Science, Rehovot 76100, Israel. Fax: +972 89 344131..ac.il, Noyame@gmail.com (N. Meital).

1. Introduction

Maximum resolution of the visual input is attained by therepeated redirection of the eyes at the visual scene. Thus,outside laboratory settings a single saccade is not performedin isolation, but rather is generated and executed in sequences.

Extensive training can lead to a specific representation of the

trained movement sequence as a unit, rather than modular

movement components (Rozanov et al., 2010; Sosnik et al., 2004;

Viviani and Cenzato, 1985); consecutive movements can com-

bine through practice into new complex movements, enlarging

the motor vocabulary.

b r a i n r e s e a r c h 1 5 2 8 ( 2 0 1 3 ) 4 2 – 4 8 43

It has been proposed that similar neural mechanisms sub-serve procedural learning in various motor systems (Doyon andBenali, 2005; Hikosaka et al., 2002; Karni et al., 1998). Hence, theconjecture that different motor systems would show comparablepractice benefit with similar time-course characteristics. A num-ber of recent studies, using an oculomotor version of the SerialReaction Time paradigm (SRT; Nissen and Bullemer, 1987),reported improvements in saccade latencies through trainingwithin session (Grosbras et al., 2001; Kawashima et al., 1998;Marcus et al., 2006). Furthermore, Albouy et al. (2006, 2008) havedemonstrated the emergence of additional between-sessiongains. However, in oculomotor SRT reactive saccades are gener-ated in pre-defined intervals, reducing the opportunity forgenerating coherent multi-element movement routines.

Here we studied the time-course of oculomotor sequencelearning to test whether saccades and fixations are optimizedthrough training, and specifically whether experience-drivenchanges in performance evolved not only within-session butalso between-sessions in a consolidation phase.

Thirteen naïve university students were trained during twoconsecutive days (Fig. 1a) to perform a sequence of SuccessiveVolitional Saccades (SVS) in a counter-clockwise direction.Saccades were aimed to four target positions; with a letterpresented in each position either a T or an L (Fig. 1b). The targetletters (array) were presented simultaneously on the screen,located in fixed positions throughout the experiment. Eye-movement data (i.e., saccades and fixations) from the partici-pants′ right eye were recorded with an Eye Link 1000 (SR-Research, Canada) eye-tracker (infrared video based system).Performance was assessed across three Time-Points computedas the mean of two successive training blocks (i.e., initialperformance blocks: 1–2, end of training blocks: 11–12, and 24 hpost-training blocks: 13–14) delineating the two study phases.

2. Result

Training on the oculomotor sequence resulted in robust gains inspeed and accuracy. Overall, practice-related improvements in

Fig. 1 – (a) The study design. (b) The target letters array.The lengths of the segments (A–D) are shown in cm and visualangles, the arrows indicate the required movement direction.

speed were evident in reduced total trial duration, reflecting animproved performance of both saccades and fixations. Accuracyimprovements were reflected in a reduction of fixation-position-offsets for the target locations (AOI) as well as areduction in the number of AOI fixation count. However, gainsin saccade parameters (i.e., duration and velocity) evolvedmainly during the post-training consolidation phase and notduring the training session.

Fig. 2a–d illustrates performance changes over the timecourse for the total trial duration. A significant decrease in trialduration was found over sessions, F(2, 622)¼8.02, p¼ .001,(Fig. 2a). However, paired t-tests comparing changes across thetwo intervals (within session, between sessions) showed thatthe reduced trial duration evolved only during the 24 h intervalfollowing the practice session (t(311)¼1.39,ns; t(311)¼2.98,p¼ .003, within and between sessions, respectively). In addition,the total trial trajectory was lengthened through training,F(2, 310)¼17.70, p¼ .000. This lengthening occurred within thetraining session, t(311)¼�4.36, p¼ .000 and was maintained onthe following day, t(311)¼�.44, ns (Fig. 2b).

As an indication of increasing accuracy, the number ofvisited AOI (number of AOIs in which at least one fixationoccurred during the trial period) significantly increased withpractice, F(2, 622)¼34.75, p¼ .000 (Fig. 2c). Fixations wereinitially less accurate, leading to less than four visited AOIs(i.e., missing the AOIs center by more than 1.51). However, asignificant increase, almost to the maximum four AOIs, wasfound within the training session, t(311)¼�4.19, p¼ .000, butalso between-sessions, t(311)¼�4.76, p¼ .000. Given the SVSdesign, the increase in the number of visited AOI, within andbetween training sessions, reflects for more accurate targetoriented saccades, thus, support the notion that answerswere not given based on previous expectations. In addition,accuracy gains were evident in the decreasing AOI fixationcount, i.e., the total number of fixations falling at each AOIper trial (number of corrective saccades) F(2, 2452)¼9.27,p¼ .000. This decrease was found in the 24 h interval follow-ing the practice session, t(1244)¼3.77, p¼ .000, but not within-session, t(1226)¼ .32, ns (Fig. 2d). These findings imply that areduced number of saccades were required to attain on-targetfixations.

The amount of time that gaze was maintained in the fourAOIs (i.e., fixation durations) decreased, F(2, 2450)¼54.02,p¼ .000, (Fig. 3a). Fixation duration decreased within thepractice session but also during the subsequent 24 h interval(t(1225)¼6.11, p¼ .000; t(1244)¼4.83, p¼ .000, respectively).

The decrease in Fixation duration was accompanied by areduction in the Fixation Offset, i.e., the distance between thecenter of the AOI and the actual fixation position (Fig. 3b).Repeated measures ANOVAs for Fixation Offsets in the X andY coordinates, over the three time-points, showed no effectover the x-axis, F(2, 1826)¼2.56, ns, but a significant decreasein the y-axis offsets with time, F(2, 1826)¼32.04, p¼ .000. They-axis offsets decreased both within the training session, t(1225)¼3.52, p¼ .000, but also during the subsequent 24 hinterval, t(1244)¼5.10, p¼ .000 (Fig. 3b).

Saccade durations measured at the three time-points,decreased, F(2, 2438)¼11.27, p¼ .000 (Fig. 3c). However, a sig-nificant improvement occurred only during the 24 h intervalfollowing the practice session (t(1219)¼�1.47, n.s; t(1239)¼4.88,

Fig. 2 – Time and accuracy measures for the total trial interval in the three study time points (initial, end and 24 h post-training). (a) Trial duration (ms), (b) trial trajectory (visual angle), (c) number of visited AOI, and (d) AOI fixation count. Errorbars indicate the standard error of the mean. npo.05, nnpo.01.

Fig. 3 – Time and accuracy measures for fixations and for saccades in the three study time points (initial, end and 24 h post-training). (a) fixation duration (ms), (b) fixation offset (visual angle), (c) saccade duration (ms), and (d) saccades velocity (deg/s).Error bars indicate the standard error of the mean. npo.05, nnpo.01.

b r a i n r e s e a r c h 1 5 2 8 ( 2 0 1 3 ) 4 2 – 4 844

p¼ .000, within and between sessions respectively). A compli-mentary rm-ANOVA for saccade velocity showed a significantincrease in average velocity across the three time-points, F(2,2438)¼27.08, p¼ .000 (Fig. 3d). There was no significant increasein velocity within the training session, nonetheless, a significantincrease in velocity occurred overnight (t(1219)¼�.39, n.s; t(1239)¼�6.58, p¼ .000, respectively).

Additional analyses were made to test whether the train-ing resulted in more efficient visual processing as reflected inthe percentage of correct answers to the question followingeach trial. The percentage of correct answers for all threesession was above 92% (M¼ .92, .97, .96, respectively). An

increase in the percentage of correct answers was observedonly within the training session but not between sessions(t(287)¼�4.11, p¼ .000; t(287)¼ .82, ns, respectively).

3. Discussion

The current results show that a new oculomotor skill, learn-ing an eye movement sequence, can be acquired throughtraining. Robust gains, in terms of movement time, speed andaccuracy, were expressed online, within the training session,as well as offline, by 24 h after the practice session. These

b r a i n r e s e a r c h 1 5 2 8 ( 2 0 1 3 ) 4 2 – 4 8 45

delayed gains indicate a distinct skill consolidation phase.Thus, the results support the notion of common basic neuralmechanisms for procedural memory consolidation under-lying the acquisition of skill in all adult brain systems(Karni, 1996; Karni et al., 1998). The current results also showthat parameters of both, saccades and fixations, can benefitfrom practice. However, perhaps because of the pre-existinghigh-efficiency of saccades, originated at various start posi-tions and notwithstanding the novel target array, withintraining session gains were more pronounced for fixationdurations, while speed and duration of saccades decreasedonly during the between-sessions interval.

The expression of speed gains with no loss of accuracyindicates skill acquisition by the ocular motor control system(Stelmach, 1996). Within the training session the total trialtrajectory was lengthened while the number of visited targetpositions (AOI) increased. The number of visited AOI as wellas the number of AOI fixations neared optimum by 24 h post-training. The co-occurrence of an increase in visited AOIs anda decrease in the AOI fixation count, suggests that trainingenhanced the knowledge of relevant AOIs and resulted in thegeneration of more accurate saccades toward these targets.However, these combined effects were not expressed by theend of the session but rather emerged within the 24 h post-training interval. The observed changes in the oculomotortrajectory may reflect the system′s solution for the high-accuracy demands of the task, but with an imposed timeconstraint. A longer movement path may enable an improve-ment in fixation positioning (better on-target accuracy) withno loss in speed (Sosnik et al., 2004). Indeed, significant gainsin speed were apparent at 24 h post-training; trial durations(i.e., sequence completion times) were significantly shor-tened, indicating a reduced execution time of the sequenceas a whole. The delayed expression of gains in trial durationmay be related to the notion of different neural mechanismunderlying the representation and processing of speed andaccuracy, with gains in the latter appearing relatively earlier(Hikosaka et al., 2002; Penhune and Steele, 2012).

The finding that the adult oculomotor system can signifi-cantly improve its performance of a pre-specified sequence ofmovements through practice, is in line with the findings instudies on the acquisition of skill in the performance of fingermovement sequences (Karni et al., 1995, 1998; Korman et al.,2003; Rozanov et al., 2010) for hand-writing like sequencelearning (Sosnik et al., 2004) and sequences of pointingmovements (Sela and Karni, 2012). Furthermore, as reportedfor manual motor sequence learning (Karni et al., 1995, 1998;Korman et al., 2003), the current results show that theacquisition of an oculomotor sequence follows two distinctlearning phases, presumably reflecting qualitative and quan-titative changes in neural representation (Doyon and Benali,2005; Hikosaka et al., 2002; Karni et al., 1998; Korman et al.,2003). Practice related gains in oculomotor performanceevolved both online, concurrently with practice, but alsooffline – after the termination of the training session. Online,within session, gains were proposed to reflect the selection,tuning and testing (by trial and error elimination) of a motorprogram from an available repertoire (Adi-Japha et al., 2008;Karni et al., 1995, 1998). Offline, delayed, gains were proposedto reflect memory consolidation processes and the generation

of lasting changes in the motor repertoire (Fischer et al., 2002;Karni et al., 1995, 1998; Korman et al., 2003; Maquet et al.,2003; Robertson et al., 2004; Robertson, 2009; Sosnik et al.,2004; Walker, 2005; Wright et al., 2010). There is a goodsupport for the notion that the expression of “delayed”,offline-gains, with no additional training in the interval,corresponds to the procedural memory consolidation periodunderlying the acquisition of motor skills (Karni, 1996;Xu et al., 2009).

The SVS task is based upon the paradigm of hand writing-like movement sequence learning (Sosnik et al., 2004, 2007),in which multiple movement elements were shown to be co-articulated into coherent multi-element movement routinesfollowing practice. The performance gains in the SVS,reported here, may reflect changes in the syntactic structureof the movement; expressed in the generation of a more tasktuned, unified representation of the motor routine as pro-posed following hand movement training paradigms(Korman et al., 2003; Rozanov et al., 2010; Shea et al., 2006;Shmuelof and Krakauer, 2011; Sosnik et al., 2004).

Albouy et al. (2006, 2008) have previously reported offline,time dependent gains in saccade latencies; while latencies ofboth trained and untrained sequences decreased at 30 minpost-training, gains were only evident for the trainedsequence on the following day. However, because the taskwas a version of the SRT paradigm, reactive saccades weretriggered in pre-defined intervals rather than concatenated togenerate a coherent multi-element movement routine. Further-more, using the SVS paradigm the current study extendsprevious findings concerning gains in saccades latencies(Albouy et al., 2006; Grosbras et al., 2001; Marcus et al., 2006),by showing a significant improvement in sequence perfor-mance as measured both by saccade and fixation parameters,and importantly by providing evidence supporting the notionthat fixations and even saccades can be optimized, albeit, indifferent phases of skill evolution.

It has been proposed that during fixations visual inputs areprocessed to optimize the subsequent eye movement (Leighand Zee, 1991). Although, several studies suggest that youngadults′ letter discrimination is very efficient and difficult toimprove within a single session of training (Dosher and Lu,2006), we cannot rule out that the reduced fixation durationswere due in part to more efficient visual processing. We proposehowever, that the reduction in fixation durations reflects areduction in motor planning demands due to the accumulatingexperience. This notion is supported by the finding that whileno gains occurred in the percentage of correct responsesbetween sessions, there were significant offline gains in fixationdurations; supporting the notion of consolidation phase gains inthe motor aspects of the SVS task. Furthermore, given thatparticipants were all highly proficient readers (university stu-dents), one may argue that even the small increases in responseaccuracy during the training session can less plausibly beascribed to within-session gains in letter recognition, than tobetter fixation positioning.

Further support for improved oculomotor planning is pro-vided by the observed reduction in fixation offset; althoughfixations were shorter, the subsequent saccades were aimedmore precisely to the targets′ locations. Given that on-targetfixation accuracy was increased, the data suggest that an

b r a i n r e s e a r c h 1 5 2 8 ( 2 0 1 3 ) 4 2 – 4 846

important aspect of learning was related to acquired, implicitknowledge of the target array. However, a significant decreasein the fixation offsets was observed only in the y-axis (verticaloffsets). Standard reading (in orthographies that are written inhorizontal text lines) requires the execution of horizontalsaccades (Engbert et al., 2005; Rayner, 1998) while relativelylarge horizontal offsets are tolerated via parafoveal reading(Rayner, 1998). Thus, horizontal offsets of fixation are highlycompensated in skilled reading and the absence of gains overthe x-axis is to be expected. There is, however, less tolerance foroffsets in the vertical axis (y-axis) which can lead to betweenline errors (Rayner, 1998). Because the target letters were notpresented on a given horizontal line, the SVS task maynecessitate the setting up of a new eye movement routine tominimize offsets over the y-axis. A similar notion about thedelayed emergence of new handmovement routines for specificsequences of hand and finger movements has been suggested(Korman et al., 2003; Shmuelof and Krakauer, 2011; Sosnik et al.,2004).

Gains in saccade duration and saccade velocity werefound only at the 24 h post-training session, i.e., after timeand perhaps a night′s sleep (Korman et al., 2003, 2007). Thedecrease in saccade duration was accompanied by increasedsaccade velocity, reflecting improvements in oculomotorparameters of performance. Given the gains in the accuracyof saccades′ landing positions (i.e., reduction of fixationoffsets) within the training session and the improvement inthe rates of correct responses, it is not likely that the (offline)delayed expression of gains in saccades′ speed were due toend of session fatigue (Adi-Japha et al., 2008).

Altogether, our results show that the reduced executiontime (saccades) was paralleled by shorter sequence planningand preparation time (fixations), indicating the acquisition ofskill in the adult oculomotor system. The larger gains infixation offsets and fixation durations compared to the gainsin saccade performance are not surprising given the pre-existing high-efficiency of saccades. Saccades are the mostrapid movements of the motor control system with presum-ably no voluntary control over the speed aspect (Leigh andZee, 1991). Practice related gains in fixation parameters werefound early as well as late in training sessions; gains insaccades parameters evolved mainly during the post-training,consolidation phase, and were expressed on the following day.Previous studies have suggested that when movements are tosome extent skilled, performance changes will evolve mainlybetween training sessions (Hauptmann and Karni, 2002;Korman et al., 2003). Therefore the differences in time-coursein the present study may relate to the level of expertise incomplex eye movement sequences attained before training.

4. Experimental procedures

4.1. Participants

Thirteen naïve university students (10 female; 23–31 yearsold) participated in the experiment. All participants wereright handed, native Hebrew speakers, with normal visionand no history of neurological disorders or chronic use ofmedications. Participants gave their written informed

consent and were paid for participation. The study wasapproved by the University of Haifa human experimentationethics committee.

4.2. Study design

The study included two sessions 24 h apart. Fig. 1a illustratesthe overall design of the study. At Day-1 participants under-went a training session (12 practice blocks; each blockcomprised 24 trials). There were mandatory 4 min breaksbetween blocks and a 10 min long break midway in thepractice session, dividing the session into two equal intervals,initial and end. Only during the long break participants wereallowed to leave the dimmed light conditions of the labora-tory. On the second day, oculomotor performance wasassessed using two blocks of trials under conditions identicalto those of Day-1. Before the practice session participantswere given 24 ‘warm-up’ trials in identical task conditions butwith a time limitation of 800 ms for completing the tasksequence. The warm up trials were given since pilot studieshave indicated that the task cannot be performed within areasonable time limit based on verbal instructions alone;subjects needed to have an implicit experience as well asan explicit notion of the location of the targets. Howevergiven the 800 ms time constraint for sequence completion,participants were exposed to the targets′ array but for mostwere unable to complete the task consistently. Oculomotorperformance under the warm up trails condition was re-testedimmediately after the session and before the Day-2 session(results will be discussed elsewhere).

Participants were seated at a distance of 50 cm in front of acomputer monitor screen (LG Flatron, L226WTQ, 22 in. diag-onal, resolution 1440�900, 16 bit, 75 Hz). A chin and foreheadrest ensured that the eye-to-screen distance was kept con-stant and that head movements were reduced to a minimum.The index and middle finger of the right hand were placed onthe numbers pad of a computer keyboard.

4.3. Task and stimuli

The task required the execution of a sequence of SuccessiveVolitional Saccades (SVS) aimed at four target positions; witha letter presented in each position either a T or an L (Fig. 1b).The target letters (white color on a black background, .32� .41visual angels) were presented simultaneously on the screen,located in fixed positions throughout the experiment. Giventhe letter positions and the eye to screen distance, not morethan one target letter could be seen during any fixation on agiven target position, making the complete target arrayimperceptible.

Each block was preceded by a standard 9-point grid calibra-tion procedure (http://www.sr-research.com/). Trials began witha drift check, ensuring fixation on the start-point (the letter X)positioned in the lower right side of the screen. Trials wereinitialized manually by the experimenter. Participants wereinstructed to scan the target array in a counter-clockwisedirection after hearing an auditory ‘go’ signal. The letter X atthe start-point was replaced either by the letter T or L after afixed delay of 300ms, redefining it as the end-point.

b r a i n r e s e a r c h 1 5 2 8 ( 2 0 1 3 ) 4 2 – 4 8 47

Trials ended with the appearance of a question after thedetection of a fixation at the end point position (i.e., the fourthtarget letter). Thus, the trigger for the termination of the trial andthe presentation of the question was contingent on fixationwithin the 1.51 radius of the fourth target letter. An interval of100ms was afforded between the time of detection of a fixationat the end point position and the presentation of the question, toexpedite the recognition of the fourth letter. Participants wererequired to answer which of the letters, T or L, appeared morethan twice in the target array. Participants responded by pressingkeyboard buttons with their right hand. For each trial, letters(T or L) were assigned to each target location randomly.To further encourage fixation at all target letters a few (2–4 perblock) catch trials contained an equal number of T and L letters;these trial were not included in the analyses. A re-calibrationwas initiated prior to the beginning of a new trial if the detectedgaze position deviated from the initial fixation point by morethan 11.

4.4. Eye-tracker measurements

Eye-movement data were recorded with an Eye Link 1000(SR-Research, Canada) eye-tracker (infrared video based system).Gaze position data was acquired at a sampling rate of 1000 Hzwith accuracy in the range of .25–.51 (manufacturer information).Binocular viewing was afforded; however, only data from theparticipants′ right eye were recorded. The experiment wasdesigned and run using the Experiment Builder software package(SR-Research, Canada).

4.5. Data analyses

Initial data processing was accomplished using the DataViewer software package (SR-Research, Canada). Measuresof time and accuracy were analyzed for the overall trial interval,saccades and fixations. Fixations were assessed using four areasof interest (AOI) with a diameter of 31 centered on the targetletters (to reflect foveal /parafoveal vision of targets).

Overall trial Duration (ms) was defined from the detectionof a fixation at the start-point position until the detection offixation at the fourth target letter. The total Trial Trajectorywas defined as the total path traveled by gaze (thus includingcorrective saccades) during the trial duration. The Number ofVisited AOI was defined as the number of AOI in which atleast one fixation occurred during the trial period. The AOIFixation Count was defined as the total number of fixationsfalling within each of the AOIs per trial.

Fixation parameters included the fixation Duration (ms).The fixation Offsets were defined as the distance between thecenter of the AOI and the actual fixation position. Fixationoffsets over start-point position, however, were not includedin the analysis, due to the drift correction procedure. Sac-cades were analyzed for Duration (ms) and Velocity (deg/sec).

Trials that contained blinks or initiation times longer than300 ms were discarded. In the case of corrective saccades(i.e., re-visiting a target point) the last fixation at the relevantAOI and the subsequent saccade were used for analysis; thelogic of this choice was that the last fixation at the relevantAOI and the subsequent saccade reflected planning and

implementation of the upcoming movement towards thesubsequent target location.

Repeated-measures ANOVAs were used to test perfor-mance across three Time-Points (i.e., initial performance, endof training and 24 h post-training) delineating the two studyphases. Each Time-Point was computed as the mean of twosuccessive blocks (blocks 1–2, 11–12 and 13–14, for initialperformance, end of training and 24 h post-training, respectively).Sphericity violations were corrected by the Greenhouse-Geisser method. Paired-samples t-tests (Bonferroni methodcorrected) were used for comparing the differences betweenthe Time-Points. Comparisons of performance parameters atinitial performance and end of training were used to assesswithin-session learning. Comparisons of performance para-meters at end of training and 24 h post-training were used toassess between-sessions, consolidation phase, gains.

Conflict of interest

The authors declare no competing financial interests.

r e f e r e n c e s

Adi-Japha, E., Karni, A., Parnes, A., Loewenschuss, I., Vakil, E.,2008. A shift in task routines during the learning of a motorskill: group-averaged data may mask critical phases in theindividuals′ acquisition of skilled performance. J. Exp. Psychol.Learn. Mem. Cognit. 34, 1544–1551.

Albouy, G., Ruby, P., Phillips, C., Luxen, A., Peigneux, P., Maquet, P.,2006. Implicit oculomotor sequence learning in humans: timecourse of offline processing. Brain Res. 1090, 163–171.

Albouy, G., Sterpenich, V., Balteau, E., Vandewalle, G., Desseilles, M.,Dang-Vu, T., Darsaud, A., Ruby, P., Luppi, P.H., Degueldre, C.,Peigneux, P., Luxen, A., Maquet, P., 2008. Both the hippocampusand striatum are involved in consolidation of motor sequencememory. Neuron 58, 261–272.

Dosher, B.A., Lu, Z.L., 2006. Level and mechanisms of perceptuallearning: learning first-order luminance and second-ordertexture objects. Vision Res. 46, 1996–2007.

Doyon, J., Benali, H., 2005. Reorganization and plasticity in theadult brain during learning of motor skills. Curr. OpinionNeurobiol. 15, 161–167.

Engbert, R., Nuthmann, A., Richter, E.M., Kliegl, R., 2005. SWIFT: adynamical model of saccade generation during reading.Psychol. Rev., 112; 777–813.

Fischer, S., Hallschmid, M., Elsner, A.L., Born, J., 2002. Sleep formsmemory for finger skills. Proc. Natl. Acad. Sci. USA 99,11987–11991.

Grosbras, M.H., Leonards, U., Lobel, E., Poline, J.B., LeBihan, D.,Berthoz, A., 2001. Human cortical networks for new andfamiliar sequences of saccades. Cereb. Cortex 11, 936–945.

Hauptmann, B., Karni, A., 2002. From primed to learn: thesaturation of repetition priming and the induction of long-term memory. Brain Res. Cogn. Brain Res. 13, 313–322.

Hikosaka, O., Nakamura, K., Sakai, K., Nakahara, H., 2002. Centralmechanisms of motor skill learning. Curr. Opinion Neurobiol.12, 217–222.

Karni, A., 1996. The acquisition of perceptual and motor skills: amemory system in the adult human cortex. Brain Res. Cogn.Brain Res. 5, 39–48.

Karni, A., Meyer, G., Jezzard, P., Adams, M., Turner, R.,Ungerleider, L., 1995. Functional MRI evidence for adult motor

b r a i n r e s e a r c h 1 5 2 8 ( 2 0 1 3 ) 4 2 – 4 848

cortex plasticity during motor skill learning. Nature 377,155–158.

Karni, A., Meyer, G., Rey-Hipolito, C., Jezzard, P., Adams, M.M.,Turner, R., Ungerleider, L.G., 1998. The acquisition of skilledmotor performance: fast and slow experience-driven changesin primary motor cortex. Proc. Natl. Acad. Sci. USA 95, 861–868.

Kawashima, R., Tanji, J., Okada, K., Sugiura, M., Sato, K.,Kinomura, S., Inoue, K., Ogawa, A., Fukuda, H., 1998.Oculomotor sequence learning: a positron emissiontomography study. Exp. Brain Res. 122, 1–8.

Korman, M., Doyon, J., Doljansky, J., Carrier, J., Dagan, Y., Karni,A., 2007. Daytime sleep condenses the time course of motormemory consolidation. Nat. Neurosci. 10, 1206–1213.

Korman, M., Raz, N., Flash, T., Karni, A., 2003. Multiple shifts inthe representation of a motor sequence during the acquisitionof skilled performance. Proc. Natl. Acad. Sci. USA 100,12492–12497.

Leigh, R.J., Zee, D.S., 1991. In: The Neurology of Eye Movementssecond ed. F.A. Davis Company, Philadelphia, PA.

Maquet, P., Laureys, S., Perrin, F., Ruby, P., Melchior, G., Boly, M.,Vu, T.D., Desseilles, M., Peigneux, P., 2003. Festina lente:evidences for fast and slow learning processes and a role forsleep in human motor skill learning. Learn. Mem. 10, 237–239.

Marcus, D.J., Karatekin, C., Markiewicz, S., 2006. Oculomotorevidence of sequence learning on the serial reaction time task.Mem. Cognit 34, 420–432.

Nissen, M.J., Bullemer, P., 1987. Attentional requirements oflearning: evidence from performance measures. Cogn.Psychol. 19, 1–32.

Penhune, V.B., Steele, C.J., 2012. Parallel contributions ofcerebellar, striatal and M1 mechanisms to motor sequencelearning. Behav. Brain Res. 226, 579–591.

Rayner, K., 1998. Eye movements in reading and informationprocessing: 20 years of research. Psychol. Bull. 124, 372–422.

Robertson, E.M., Pascual-Leone, A., Miall, R.C., 2004. Currentconcepts in procedural consolidation. Nat. Rev. Neurosci. 5,576–582.

Robertson, E.M., 2009. From creation to consolidation: a novelframework for memory processing. PLoS Biol. 7, e1000019.

Rozanov, S., Keren, O., Karni, A., 2010. The specificity of memoryfor a highly trained finger movement sequence: change theending, change all. Brain Res., 1331; 80–87.

Sela, I., Karni, A., 2012. Differences in learning volitional (manual)and non-volitional (posture) aspects of a complex motor skillin young adult dyslexic and skilled readers. PLoS One 7,e43488.

Shea, C., Park, J., Braden, H., 2006. Age-related effects insequential motor learning. Phys. Ther. 86, 478–488.

Shmuelof, L., Krakauer, J., 2011. Are we ready for a natural historyof motor learning? Neuron 72, 469–476.

Sosnik, R., Flash, T., Hauptmann, B., Karni, A., 2007. Theacquisition and implementation of the smoothnessmaximization motion strategy is dependent on spatialaccuracy demands. Exp. Brain Res. 176, 311–331.

Sosnik, R., Hauptmann, B., Karni, A., Flash, T., 2004. Whenpractice leads to co-articulation: the evolution ofgeometrically defined movement primitives. Exp. Brain Res.156, 422–438.

Stelmach, G., 1996. Motor learning: toward understandingacquired representation. In: Bloedel, J., Ebner, T., Wise, S.P.(Eds.), The Acquisition of Motor Behavior in Vertebrates. MITPress, Cambridge, MA, pp. 391–408.

Viviani, P., Cenzato, M., 1985. Segmentation and coupling incomplex movements. J. Exp. Psychol. Hum. Percept. Perform.11, 828–845.

Walker, M., 2005. A refined model of sleep and the time course ofmemory formation. Behav. Brain Sci. 28, 51–104.

Wright, D.L., Rhee, J.H., Vaculin, A., 2010. Offline improvementduring motor sequence learning is not restricted to developingmotor chunks. J. Mot. Behav. 42, 317–324.

Xu, T., Yu, X., Perlik, A.J., Tobin, W.F., Zweig, J.A., Tennant, K.,Jones, T., Zuo, Y., 2009. Rapid formation and selectivestabilization of synapses for enduring motor memories.Nature 462, 915–919.