visually-guided correction of hand reaching movements: the neurophysiological bases in the cerebral...

1

3

4

5

6

7 Q1

89

10

111213

1 5

16171819

2021222324252627

2 8

46

47

48

49

50

51

52

53

54

55

56

57

58

59

60

61

Vision Research xxx (2014) xxx–xxx

VR 6900 No. of Pages 13, Model 5G

30 September 2014

Contents lists available at ScienceDirect

Vision Research

journal homepage: www.elsevier .com/locate /v isres

Visually-guided correction of hand reaching movements:The neurophysiological bases in the cerebral cortex

http://dx.doi.org/10.1016/j.visres.2014.09.0090042-6989/� 2014 Published by Elsevier Ltd.

⇑ Corresponding author at: Department of Physiology and Pharmacology, SAPI-ENZA University of Rome, Piazzale Aldo Moro 5, 00185 Rome, Italy. Fax: +39 064991 0942.

E-mail address: [email protected] (R. Caminiti).1 All authors contributed equally to the manuscript.

Please cite this article in press as: Archambault, P. S., et al. Visually-guided correction of hand reaching movements: The neurophysiological basecerebral cortex. Vision Research (2014), http://dx.doi.org/10.1016/j.visres.2014.09.009

P.S. Archambault a,b,1, S. Ferrari-Toniolo c,1, R. Caminiti c,⇑,1, A. Battaglia-Mayer c,1

a School of Physical & Occupational Therapy, McGill University, Montreal, Canadab Interdisciplinary Research Centre in Rehabilitation, Montreal, Canadac Department of Physiology Pharmacology, SAPIENZA University of Rome, Rome, Italy

a r t i c l e i n f o

293031323334353637383940

Article history:Received 13 March 2014Received in revised form 18 September 2014Available online xxxx

Keywords:Visual control of reachingHand trajectory formation and correctionParietal cortexMotor cortexPremotor cortexBehavioral neurophysiology

41424344

a b s t r a c t

The ability of human and non-human primates to make fast corrections to hand movement trajectoriesafter a sudden shift in the target’s location is a key feature of visuo-motor behavior. In healthy individ-uals, hand movements smoothly adapt to a change in target location without needing to complete themovement to the first target location, as typical of parietal patients. This finding indicates that the ner-vous system continuously monitors the visual scene and is able to integrate new information in order toproduce an efficient motor response. In this paper, we review the kinematics, reaction times and muscleactivity observed during the online correction of hand movements as well as the underlying neurophysi-ological processes studied through single-cell neural recordings in monkeys. Brain stimulation, lesion andimaging studies in humans are also discussed. We demonstrate that while online correction mechanismsstrongly depend on the activity of a parieto-frontal network of which the posterior parietal cortex is acrucial node, these mechanisms proceed smoothly and are similar to what is observed during simplepoint-to-point movements. Online correction of hand movements would rely on feedforward and feed-back mechanisms in the parietal cortex, as part of the activity within the fronto-parietal network forthe planning and execution of visuo-motor tasks.

� 2014 Published by Elsevier Ltd.

45

62

63

64

65

66

67

68

69

70

71

72

73

74

75

76

77

1. Introduction

Reaching for objects and targets in the environment is an essen-tial aspect of behavior in both human and non-human primates.This activity requires an adequate coordination and integration ofsensory (vision, proprioception) and motor processes. Reachingfor stationary objects requires constant feedforward and feedbackactivity within the Central Nervous System (CNS). At a first level,feedforward processes are necessary to plan out the reachingmovement, while feedback is used to control for errors duringmovement execution and to monitor the outcome. Human andnon-human primates are able to adapt their ongoing movementin response to a rapid change in the target’s location. In such targetshift conditions, the CNS does not complete the hand movement tothe original target, but smoothly adjusts it in order to reach thesecond target location. This will be possible if there is enough time

78

79

80

81

82

83

for the online correction, given the natural hand movement reac-tion time. The smooth, online correction of hand movement inresponse to a shift in target location constitutes evidence thatvisual information has a continuous access to the CNS and thatbrain centers continuously monitor ongoing movements, makingthe required adjustments due to changing task demands.

In humans, lesions to the posterior parietal cortex (PPC) canlead to optic ataxia (OA), a condition characterized by the inabilityto accurately guide the hand to visual targets in the absence ofpurely motor or visual deficits. Patients with OA are unable to cor-rect their ongoing hand movements in a double-step target para-digm (Battaglia-Mayer et al., 2014; Gaveau et al., 2014; Greaet al., 2002; Pisella et al., 2000; Prablanc, Desmurget, & Grea,2003). Instead, they complete their hand movement to the originaltarget location, before redirecting their hand toward the secondtarget. The parietal, dorsal premotor and motor cortical areas arethought to form a recurrent network that is crucial for the coordi-nation of hand and eye movements, including planning, executionand control (for reviews see Battaglia-Mayer et al., 2003; Caminitiet al., 2010). The parietal areas are the sources of the visual inputwhereby this distributed network composes and controls handmovements to visual targets (Battaglia-Mayer et al., 2014;

s in the

http://dx.doi.org/10.1016/j.visres.2014.09.009

mailto:[email protected]


http://www.sciencedirect.com/science/journal/00426989

http://www.elsevier.com/locate/visres


84

85

86

87

88

89

90

91

92

93

94

95

96

97

98

99

100

101

102

103

104

105

106

107

108

109

110

111

112

113

114

115

116

117

118

119

120

121

2 P.S. Archambault et al. / Vision Research xxx (2014) xxx–xxx


30 September 2014

Caminiti, Ferraina, & Johnson, 1996; Johnson et al., 1996). The roleof these areas with respect to the online control of movement hasbeen studied over the last 30 years in both human and non-humanprimates, using a variety of methodologies including psychophys-ical measurements, transcranial magnetic stimulation (TMS),lesion studies, imaging and neural activity recordings. The objec-tive of this paper is to review and compare the evidence relatedto the cortical control of the online correction of hand movements.

122

123

124

125

126

127

128

129

130

131

132

133

134

135

136

137

138

139

140

141

142

143

144

145

146

2. Psychophysics of double-step hand movements

To study hand and eye behavior during sudden changes ofmotor plan, a ‘‘single-step/double-step’’ paradigm may be adopted.In these experiments, the subject starts with the hand or finger atan initial location, then is presented with a first target and isinstructed to reach for it as fast as possible (single-step condition).However, in some random trials, the location of the reach targetsuddenly shifts in space (double-step condition), requiring a redi-rection of the hand movement toward this new location. The ratioof single- to double-step trials, as well as their order, is kept fromthe subjects in order to prevent any prediction, that could influ-ence their behavior.

In single-step (direct) trials, there is an initial eye saccade to thetarget, followed by a hand movement. Note that the initial saccademay be followed by one or more corrective saccades. Thus we dis-tinguish an eye reaction time (eRT1), a hand reaction time (hRT1)and a hand movement time (hMT). In double-step (corrected)reaches, the presentation of the second target leads to anothereye saccade, followed by a change in the hand trajectory. Whereaseye saccades occur quickly and are not corrected mid-flight, thehand movement trajectory is adjusted online after the presentationof the second target. Thus we distinguish a first and second eye

Fig. 1. Experimental apparatus, task, and behavioral performance. Monkeys performed simovement corrections (B1) of 90� (from the center to target 5 and then to target 1), or 1where target where presented at the vertices of an imaginary cube, in an intermingled rand which trial (single- or double-step) they had to perform. Lit targets were positioned bdifferent directions of direct reach trials. (A3) Examples of eye (thin curve) and hand (thicmovement (0). The red triangle on the abscissa indicates the moment of target presentafrom 8 to 1. The hand path originally directed to target 8 reverses toward target 1, after pthat the path length toward the first target is a function of the time the target stays on, thand movement (hMT1). (B3) Hand (thick curves) and eye (thin curves) velocity profileshMT onset (0, blue curves). Triangles refer to the time of first and second target presenta(green) or at the onset of movement-time (blue). (Reproduced with modifications from

Please cite this article in press as: Archambault, P. S., et al. Visually-guided corcerebral cortex. Vision Research (2014), http://dx.doi.org/10.1016/j.visres.2014.

reaction time (eRT1, eRT2), a first and second hand reaction time(hRT1, hRT2) and a hand movement time (hMT). Behavior duringsingle- and double-steps reaches is summarized in Fig. 1.

For an online correction to be observed, there needs to beenough time for the subject to react; e.g., the interval betweenthe movement time to the first target and the time of shift in targetlocation should be larger than the visuo-motor reaction time. Con-versely, if the shift in target occurs too early (e.g., at the start of theinitial reaction time), then the subject will reach for the second tar-get directly, without initiating a movement toward the first target.Conversely, if the shift occurs too late, then the movement to thefirst target may already have been completed. Thus, in general,the shift in target location ranges from the later part of the initialhand reaction time period, to some maximum time during handmovement, depending on extent and speed. Within these timingconstraints, subjects are able to produce smooth, corrective handmovements in response to a shift in target (Oostwoud Wijdenes,Brenner, & Smeets, 2011).

Paradigms of double-step hand movements presented in the lit-erature vary both in terms of location and timing of the target shift(for recent reviews see Battaglia-Mayer et al., 2014; Gaveau et al.,2014). In experiments, targets can be displaced either parallel tothe initial movement direction, e.g. further and closer (Soechting& Lacquaniti, 1983) or, more frequently, perpendicular to themovement direction (Brenner & Smeets, 1997; Briere & Proteau,2011; Desmurget et al., 1999; Goodale, Pelisson, & Prablanc,1986; Gritsenko, Yakovenko, & Kalaska, 2009; Johnson, VanBeers, & Haggard, 2002; Oostwoud Wijdenes, Brenner, & Smeets,2011; Prablanc & Martin, 1992; Proteau, Roujoula, & Messier,2009; Reichenbach et al., 2009; Veerman, Brenner, & Smeets,2008). Timing of the target shift can be set within 0–200 ms fromthe onset of hand movement (Brenner & Smeets, 1997; Briere &Proteau, 2011; Johnson, Van Beers, & Haggard, 2002; Oostwoud

ngle-step direct reaches (A1; from the center to target 8) or made double-steps hand80� (from center to target 8 and then to target 1), within a reaction-time paradigmandomized design. Therefore, animals could not predict which target would appeary two robot-arms in total darkness. (A2) Examples of hand movement trajectories ink curve) speed profile during single-step, direct reaches, aligned to the onset of handtion. (B2) Examples of corrected reaches when a change of target location occurredresentation of the latter during hRT1 (green) or at the onset of hMT1 (blue). Notice

herefore it is shorter when the target jump occurs during hRT1 than at the onset ofduring double-step reaches, when the target jumps during hRT1 (green curves) or attion in different reaching conditions, i.e. target jump occurring during reaction-timeArchambault, Ferrari-Toniolo, & Battaglia-Mayer, 2011.)

rection of hand reaching movements: The neurophysiological bases in the09.009


147

148

149

150

151

152

153

154

155

156

157

158

159

160

161

162

163

164

165

166

167

168

169

170

171

172

173

174

175

176

177

178

179

180

181

182

183

184

185

186

187

188

189

190

191

192

193

194

195

196

197

198

199

200

201

202

203

204

205

206

207

208

209

210

211

212

213

214

215

216

217

218

219

220

221

222

223

224

225

226

227

228

229

230

231

232

233

234

235

236

237

238

239

240

241

242

243

244

245

246

247

248

249

250

251

252

253

254

255

256

257

258

259

260

261

262

263

264

265

266

267

268

269

270

271

272

273

274

275

276

277

278

P.S. Archambault et al. / Vision Research xxx (2014) xxx–xxx 3


30 September 2014

Wijdenes, Brenner, & Smeets, 2011; Veerman, Brenner, & Smeets,2008) or after a pre-determined amount of hand displacement(Proteau, Roujoula, & Messier, 2009; Reichenbach et al., 2009). Inthose cases, the shift in target occurs during hand MT. Others haveused a fixed time after the presentation of the first target, such thatthe shift in target can then occur either during hand RT or MT(Soechting & Lacquaniti, 1983). Alternatively, target shift can occurduring hand RT when it is triggered with the first ocular saccade(Desmurget et al., 1999; Goodale, Pelisson, & Prablanc, 1986;Gritsenko, Yakovenko, & Kalaska, 2009; Prablanc & Martin, 1992).When this paradigm is employed, subjects can even be unawareof the target displacement, as conscious visual perception isreduced during ocular saccades. The target displacement needs tobe relatively small for this phenomenon to occur. For example,Gritsenko et al. reported that for a target displacement correspond-ing to a change in visual angle of 3.5�, participants noticed the tar-get jump in approximately 30% of trials (Gritsenko, Yakovenko, &Kalaska, 2009).

Irrespective of the double-step paradigm employed, the timingof corrective hand and eye movements is remarkably stable. Thereis a temporal synchronization between eye and hand, as the handRT has been observed to always follow the onset of the eye saccadeby 50–100 ms (Prablanc, Desmurget, & Grea, 2003), in either sin-gle- or double-step conditions, and for either the first or secondtarget. The hand correction movement time (hRT2) has beenreported in most studies to range from 100 to 200 ms (Brenner &Smeets, 1997; Briere & Proteau, 2011; Desmurget et al., 1999;Goodale, Pelisson, & Prablanc, 1986; Gritsenko, Yakovenko, &Kalaska, 2009; Oostwoud Wijdenes, Brenner, & Smeets, 2011;Prablanc & Martin, 1992; Proteau, Roujoula, & Messier, 2009;Soechting & Lacquaniti, 1983; Veerman, Brenner, & Smeets,2008), although longer latencies (300–350 ms) have also beenreported (Johnson, Van Beers, & Haggard, 2002; Reichenbachet al., 2009). It is interesting to note that hRT2, overall, does notseem to be influenced by the timing of the target jump. The studiesmentioned above used a variety of paradigms, i.e., with the targetshifting at various times during hMT or after various amounts ofhand displacements, with no noticeable changes on hRT2. Further,Briere and Proteau (2011) had the same subjects perform double-step movements where the target was shifted by either 15 or30 mm, and at either 150 or 250 ms after the onset of hand move-ment. No significant differences were found in the latencies of thecorrective hand movements for these various conditions.

It is also possible that the different latencies reported in the lit-erature might be due to variations in methodologies. Indeed, thetiming of the divergence of the hand trajectory in double-step ver-sus single-step conditions has been calculated based on position,velocity or acceleration of the hand, and either using a fixedthreshold, a relative threshold or a confidence interval, amongstothers (Oostwoud Wijdenes, Brenner, & Smeets, 2013). In addition,linear filtering methods that are commonly used with kinematicdata may introduce time shifts that anticipate the onset of move-ment (Robertson & Dowling, 2003), which may explain why somevery fast corrective movements have been reported. Through sim-ulations of hand movement trajectories, Ooswoud Wijdenes andcollaborators have proposed that an extrapolation method maybe the most reliable in order to correctly measure hRT2. Thismethod involves identifying a portion of the double-step hand tra-jectory that clearly deviates from that of the single-step trajectory,then extrapolating back in time to find the onset of this deviation.

Tangential velocity of hand and eye movements for single-steptasks follows a bell-shaped profile (Fig. 1A3). In double-step tasks,two successive bell-shaped profiles are observed (Fig. 1B3). Peakhand velocity is generally higher for the corrected movement indouble-step tasks, than for the movement to the first target, orfor single-step movements. Despite this higher velocity, being the


hand trajectory in double-step tasks usually longer than in sin-gle-step (Fig. 1B1–B2), the total movement time results to be alsolonger during corrected movement (Fig. 1B3), as compared to directones (Fig. 1A3).

Other work has indicated that not only the appearance or tim-ing of targets can influence motor behavior, but also the nature ofthe stimulus and the task demands. For example, in blocks of trialswhere the visual representation of the moving finger is suddenlydisplaced, but returns to its accurate location, subjects learn to dis-miss this visual perturbation (Franklin & Wolpert, 2008). Knill,Bondada, and Chhabra (2011) have used vertical and horizontalrectangles as targets in a reaching task where the visual represen-tation of the finger was suddenly shifted. While the timing of thecorrective response remained the same, the amplitude of the cor-rective movement changed with target shape: correction wasmuch less when the perturbation occurred in the same dimensionas the long axis of the target rectangle; i.e., for a vertically-orientedtarget, a vertical perturbation of the finger’s representation eliciteda smaller corrective response than for a horizontal target, wheremovement constraints are more demanding. Likewise, otherresearch has shown that accuracy to a final target, after a suddendisplacement of the finger’s representation, is improved when sub-jects are allowed to hit the target rather than to stop on it (Liu &Todorov, 2007). Presumably, when subject are allowed to hit thetarget, stability constraints are reduced at the benefit of accuracy.These results illustrate that the nervous system is able to modulatefeedback processes in order to face task demands, an idea that iscentral to the optimal feedback control theory of motor coordina-tion (Todorov & Jordan, 2002).

Aging significantly lengthens the latency of hRT2 in double-stepconditions (Kadota & Gomi, 2010; Sarlegna, 2006). However, stud-ies disagree on the extent of this age-related change. Sarlegnareported important differences due to aging, with their older groupsometimes almost reaching the first target in double-step condi-tions before correcting their hand moment (Sarlegna, 2006). Onthe other hand, Kadota and Gomi found significant but small differ-ences in hRT2 between their older and younger subjects. Interest-ingly, this difference was much smaller than the age-relateddifference in a visual discrimination task, where the same subjects,in other trials, had to press a button as soon as they perceived achange in target location (Kadota & Gomi, 2010). One possibleexplanation for the differences in the two studies may be move-ment speed. In Kadota and Gomi’s (2010), the required hand move-ment was 30 cm, and subjects were instructed to move in 0.6 s,timed by a metronome. By contrast, in Sarlegna’s (2006), subjectshad to reach 15 cm in 1 s, or almost twice as slow. It could be thatcorrective hand movements are faster at higher speeds.

Only a few studies have looked at reaction times for double-steptasks in monkeys. Similar to trials in humans, Georgopoulos,Kalaska, and Massey (1981) reported a hRT1 of about 220 ms withno appreciable variations for hRT2, although a trend for the latter tobe a bit longer was observed in the case when the second target waspresented shortly after (50–100 ms) the first one. Archambault,Caminiti, and Battaglia-Mayer (2009) and Archambault, Ferrari-Toniolo, and Battaglia-Mayer (2011) reported average eye RT’s of200 ms for two monkeys, with no differences between the firstand second target in double-step trials. Hand reaction times were325 and 265 ms, for hRT1 and hRT2 respectively. By contrast(Dickey, Amit, & Hatsopoulos, 2013) used a modelling and extrapo-lation procedure based on the hand velocity profiles, and estimatedaverage hRT1 and hRT2 of approximately 200 ms, in three monkeys.While the differences between these sets of findings may beexplained by the methodology employed in calculating hand RT’s,the values reported for all the primate studies are generally higherthan those that have been reported for human subjects. This is sur-prising, considering that hand RTs to visual targets are generally



279

280

281

282

283

284

285

286

287

288

289

290

291

292

293

294

295

296

297

298

299

300

301

302

303

304

305

306

307

308

309

310

311

312

313

314

315

316

317

318

319

320

321

322

323

324

325

326

327

328

329

330

331

332

333

334

335

336

337

338

339

340

341

342

343

344

345

346

347

348

349

350



30 September 2014

faster in monkeys than humans. One partial explanation is thatmonkeys participating in behavioral experiments typically receiveextensive training, with their performance dictated by a rewardmechanism. Monkeys thus naturally find ways to maximize theprobability of reward while minimizing effort. Because double-stepand single-step trials are intermingled, it is possible that monkeyslearn to wait as long as they can after the presentation of the firsttarget, or learn to move slower within the constraints of the task,in order to account for the possibility that the target may or maynot switch; i.e., by waiting longer they reduce the probability oferror (continuing to the first target in double-step conditions).

351

352

353

354

355

356

357

358

359

360

361

362

363

364

365

366

367

368

369

370

371

372

373

374

375

376

377

378

379

380

381

382

383

384

385

386

387

388

389

390

391

392

3. Muscle activity during correction of hand movements

Studies of electromyographic (EMG) activity during the onlinecorrection of hand movements first revealed that arm muscleactivity initially proceeds the same way in single-step as in dou-ble-step conditions, before the switch in targets (Gielen, van denHeuvel, & Denier van der Gon, 1984; Soechting & Lacquaniti,1983). EMG activity then diverges in double-step trials comparedto the single-step condition, at a latency of around 100 ms afterthe change in target location, and preceding the changes in handtrajectories (Fautrelle et al., 2010; Reichenbach et al., 2009;Soechting & Lacquaniti, 1983). This latency is consistent with whathas been observed for kinematic data, considering that EMG burstprecedes changes in kinematics, with a variable electromechanicaldelay depending on the current state of the limb. The patterns ofmuscle activity are consistent with the joint torques needed forgenerating the corrective movements. Thus, the CNS’s control ofmuscle activity during corrective movements seems to be quitesimilar in nature to its control during point-to-point movements.While the CNS, due to redundancy, can produce the same joint tor-ques using different combinations of muscle activities, there doesnot seem to be any special mode or pattern of muscle activationfor online control and adjustments. This was further confirmedby studies from d’Avella, Portone, and Lacquaniti (2011) who com-pared muscle synergies, or combinations of EMG patterns, duringsingle-step and double-step movements. EMG activity wasrecorded from 16 arm muscles, and synergies were computedusing an iterative optimization algorithm (d’Avella, Portone, &Lacquaniti, 2011; d’Avella et al., 2006). The same synergies wereobserved in corrected and uncorrected movements; only theirmodulation differed (how much one or another synergy is usedat a given time). Thus, the CNS does not need to use a special mus-cle synergy for online correction, but modulates existing synergiesin order to produce the required movement.

Whole body coordination likewise continues during online cor-rection, as it occurs during uncorrected reaching movements.Anticipatory postural responses are observed in the leg muscula-ture when subjects perform single- and double-step arm move-ments in standing, both during the initiation of movement andduring the correction (Fautrelle et al., 2010; Leonard et al., 2011).

393

394

395

396

397

398

399

400

401

402

403

404

405

4. Neurophysiology studies in the cerebral cortex

Georgopoulos and collaborators were the first to study singleneuron activity in monkeys performing single- and double-stephand movements (Georgopoulos et al., 1983). They recorded a totalof 79 neurons from three monkeys in motor cortex (area 4, M1).Monkeys performed a planar reaching task with a manipulandum,from a central location to eight possible targets placed in a circle.In some trials, the reach target could be extinguished after50–400 ms and unexpectedly changed to the opposing one; thustarget switch occurred either during hand reaction or hand move-ment time. They observed that cells displayed single-peaked and


double-peaked activity profiles in single- and double-step trials,respectively, corresponding to the single- and double-peaked handvelocity profiles. They also noted that the shape of cell activity forthe corrected movements in double-step trials closely resembledthat of the control condition, i.e., a single-step trial to the same tar-get. For example, a cell showing a decrease in activity (silent period)at the start of a single-step movement to a particular target woulddisplay the same drop in activity during a correcting movement tothat same target. Finally, they reported higher activity for correctedmovements of double-step trials than in the control, single-stepconditions, which they attributed to the higher hand velocity dur-ing the correction.

Twenty year later, Archambault, Caminiti, and Battaglia-Mayer(2009), Archambault, Ferrari-Toniolo, and Battaglia-Mayer (2011)studied the activity of single neurons located in the dorsal premo-tor cortex (PMd; area 6; 119 cells), M1 (area 4; 155 cells) and theposterior parietal cortex (PPC; area 5; 250 cells) in two monkeys.The paradigm required unconstrained natural reaches (no manipu-landum) in three dimensions, from a center location to one of eightpossible targets located at the vertices of an imaginary cube(Fig. 1A). In double-step trials (Fig. 1B), the reach target could beswitched either during hRT1 or at the initiation of hMT, and eitherto the opposite or to the adjacent target. They also observed dou-ble-peaked neural activity patterns in all three cortical areas andfor all double-step conditions (Fig. 2). Signalling of both target pre-sentation during direct reaches and of the future change of move-ment direction in the corrected ones (calculated as the time inwhich cell activity in the corrected reached diverged from that ofthe direct reaches) occurred first in premotor cortex. Using a linearregression approach, the relationship between cell activity andhand movement parameters (velocity, direction, target location)was studied. It was found that M1, PMd and PPC all contained cellsthat encoded hand kinematics equally well, whether the move-ment included a change in target location or not, and whether thatchange in target was to the adjacent or opposite target, or theswitch occurred during hRT1 or hMT (Archambault, Caminiti, &Battaglia-Mayer, 2009; Archambault, Ferrari-Toniolo, & Battaglia-Mayer, 2011). This allowed accurate reconstruction of ‘‘neural’’ tra-jectories that well matched the experimentally observed ones(Fig. 3A). On the basis of the observed cell activity patterns, therewere no cells that specifically, or uniquely, signalled the occur-rence of a change in target location. However, the predictive powerof neural activity decreased somewhat in the PMd and M1 cells,but not in PPC (Fig. 3B). This may be indicative of a greater roleof the PPC in the online adjustment of hand trajectory, as comparedto PMd and M1, consistent with the observation that in humanslesions in the PPC lead to optic ataxia, a deficit consisting of errorsin reach endpoint and in adjusting hand movements to rapidlyshifting targets (for a recent review see Battaglia-Mayer et al.,2014, and the reference therein). Of course, ongoing hand kinemat-ics does not completely explain neural activity and other factorsmay be at play. For example, motor cortex, thanks to the directaccess to the spinal cord, can also have a pivotal role in encodingthe force necessary to brake the hand movement to the first targetand then accelerate it toward the second one.

On the other hand, the decrease in correlation from one move-ment condition to another could be a consequence of the interfer-ence occurring when new information coexists with an originalmotor intention and calls for its change.

In recent work, Dickey, Amit, and Hatsopoulos (2013) analyzedthe activity of 75 M1 cells, 331 PMd cells and 90 ventral premotor(PMv) cells from three animals. No information was provided as tothe precise area (F5, F4) of recording in PMv. They studied 1-dimensional single-joint (elbow) movements with and without achange in target location. With their arm in a manipulandum,monkeys had to perform elbow extension movements to one of



406

407

408

409

410

411

412

413

414

415

416

417

418

419

420

421

422

423

424

425

426

427

428

429

430

431

432

433

434

435

436

437

Fig. 2. (A) Neural activity during direct and corrected reaches in PMd, M1, and PPC. Examples of single cell activity (as spike density functions; SDF) during direct (single-step)and corrected (double-step) reaches in PMd, M1 and PPC. SDFs are indicated by black curves, while the gray curves represent the hand velocity profiles. In direct reaches, theblack triangle indicates the moment of target presentation, while 0 is the hand movement onset. For correct reaches, the black and white triangles indicate first and secondtarget presentation, respectively, with 0 the onset of hand movement. On the left y-axis, units for SDF are represented in spikes/second (sp/s); on the right y-axis units refer tospeed profile in m/s. (B) Cumulative frequency distributions of the onset times of cell activity during direct reaches (left) and the times (right) when cell activity in thecorrected reaches diverged from that observed during direct reaches (‘divergence time’), after second target presentation at the onset of movement time. The three curves ineach plot refer to the different population of neurons (PMd, red; M1, green and PPC, blue). (Reproduced with modification from Archambault, Ferrari-Toniolo, & Battaglia-Mayer, 2011.)



30 September 2014

five possible targets, located at equidistant points within an arc.Double-step conditions could occur for movements to the middlethree targets, with the target shifting either forward or backwardby one position, and was triggered with the start of hMT. They alsoconfirmed, using a linear regression model, that neural activity inM1, PMd and PMv encode kinematics in both single- and double-step tasks. Their modelling of cell activity in the double-step con-ditions was based on that observed during the single-step condi-tions, in order to determine if the latter could predict the former.Indeed, Georgopoulos et al. (1983) had previously observed thatthe shape of the neural activity in PPC during double-step trialsresembles that of the corresponding single-step trials; i.e., the neu-ral activity in a particular cell for a double-step movement from acentral position to targets, let’ say, 1 and then 8 (see Fig. 1B1 andB2) seems to be a combination, on joining, of the activity observedin single-step reaching to from the center to 8, followed by the


activity for the movement from the center to 1. Based on thisobservation, Archambault et al. (2009, 2011) performed a quantita-tive analysis of cell activity by first determining the timing of thetwo peaks in the hand velocity profiles of double-step tasks. Theythen joined (added) the two neural activity profiles of single-steptrials using the same time delay, and compared this reconstructedactivity profile with that recorded during single-step trials (Fig. 4).The authors found that indeed, double-step neural activity couldbe predicted from single-step activity in M1, PMd and PPC cells,as the correlation between real and predicted neural activity washigher when using the corresponding single-step trials, rather thanrandom ones.

In the studies of Archambault et al. (2009, 2011), no assumptionwas made on how exactly the nervous system would combinesingle-step movements for double-step conditions. Dickey, Amit,and Hatsopoulos (2013) went one step further in their analysis by



438

439

440

441

442

443

444

445

446

447

448

449

450

451

452

453

454

455

456

457

458

459

460

461

462

463

464

465

466

467

468

469

470

471

472

473

474

475

476

477

478

479

480

481

482

483

484

485

486

487

488

489

490

491

492

493

494

495

496

497

498

499

500

501

502

503

504

505

506

507

508

509

510

511

512

513

514

515

516

517

518

519

520

521

522

523

524

525

526

Fig. 3. Encoding of limb kinematics during direct and corrected reaches in PMd, M1, and PPC. (A) Comparisons of real trajectories (interrupted line) performed by the monkeywith ‘‘neural’’ trajectories computed from cell activity in PMd, M1 and PPC for some examples of corrected reaches. For PMd and PPC, the trajectories correspond tomovements with correction from target 4 to 5 (with target jump occurring during hRT1), while for M1 the trajectory is first to target 7 then 2 (with target jump at the onset ofmovement-time). (B) Distribution of the Pearson correlation coefficients (r) between real and ‘‘neural’’ trajectories, during direct (left) and corrected (right) reaches, fordifferent areas. Asterisks indicate statistically significant difference (p < 0.05), n.s. non-significant ones. Notice that r values are significantly higher for direct than forcorrected reaches in both PMd and M1, but not in PPC, suggesting a special role of this area in trajectory modification. (Reproduced with modifications from Archambault,Ferrari-Toniolo, & Battaglia-Mayer, 2011.)



30 September 2014

contrasting two possible models of the control of corrective handmovements. In one approach, the authors assumed that the neuralcontrol mimicked the smooth transition of the hand kinematicsfrom one target to the other (‘‘superposition’’ hypothesis). In theother approach (‘‘replacement’’ hypothesis), the neural control sig-nal, based on hand velocity, would abruptly stop after the switch intarget and be replaced by that from the initial to the second target.Their analysis involved first determining the relationship betweenhand velocity and cell activity using single-step trials only, througha linear regression analysis. Then, they predicted the double-stepneural activity using the control signals from either the superposi-tion or replacement hypotheses. This is different than inArchambault et al. (2009, 2011), where the double-step cell activityprofiles were directly compared with joined single-step activityprofiles. Dickey, Amit, and Hatsopoulos (2013) found that the super-position hypothesis was better at explaining double-step from sin-gle-step activity for 60% of the recorded neurons in both M1 andPMd/PMv, while the replacement hypothesis had a higher predic-tive power for 40% of neurons. This would indicate the presence ofsimultaneous, parallel strategies in the CNS during the transitionbetween movements from the first to the second target. This paral-lel control does not continue during movement to the second target(see also Section 11 for further discussion on this topic).

The linear regression analyses performed by the two aforemen-tioned groups also included the calculation of a delay betweenthe activity of each recorded neuron and hand movement(Archambault, Caminiti, & Battaglia-Mayer, 2009; Archambault,Ferrari-Toniolo, & Battaglia-Mayer, 2011; Dickey, Amit, &Hatsopoulos, 2013). It was found that for a majority of cells inM1 and PMd, neural activity preceded the change in hand kinemat-ics while in PPC the activity in a majority of cells lagged hand kine-matics, although both types of cells (leading and lagging) wereseen in all areas. This was interpreted as the PPC receiving morevisual and proprioceptive feedback signals about the hand move-ment than the frontal areas. In Dickey et al.’s analysis, a greaterproportion of cells following the replacement model also laggedthe modulations in hand kinematics, compared to cells followingthe superposition hypothesis. Of course, PPC is not the sole recipi-ent of feedback information. For example, Pruszynski et al. (2011)have demonstrated that M1 neurons are able to appropriatelyadjust their level of activity in order to counter the effects of amechanical perturbation.

Additional interesting information about the timing of neuralactivity and hand trajectory correction emerges from the analysisof the population activity, since this provides clues as to under-standing when a future change in hand trajectory is signalled in


the parieto-frontal system (Archambault, Caminiti, & Battaglia-Mayer, 2009; Archambault, Ferrari-Toniolo, & Battaglia-Mayer,2011; Battaglia-Mayer et al., 2014). When the second target is pre-sented during planning of hand movement to the first target(Fig. 5), therefore during hRT1, the population activity associatedto double-step movements in all areas studied diverges from thatof single-step reaches, thus signalling the change of movement tra-jectory, before or just around the onset of hand movement towardthe first target. Therefore, within the time window elapsing fromthe divergence in neural activity and the change of movement tra-jectory, a new signal related to the preparation of the futurechange of hand movement emerges, suggesting that during thistime old and new command signals influence cell activity in theparieto-frontal system. The earlier signalling occurs before move-ment onset in PMd, the late ones in M1 and PPC. These same timerelationships are maintained across areas when the target shiftoccurs at the onset of hand movement. In this case, however, thesignal about the change in hand trajectory occurs closer to theonset of hand movement toward the first target. Therefore, thetemporal evolution of the population activities suggests a differen-tial role of the three areas in online control. The activation peaks inPMd, both before the first and second target presentation; in otherwords the signal ‘‘to go’’ is combined with the signal ‘‘to correct’’the original motor plan. M1 seems to signal hand movement initi-ation and the precise control of hand kinematics on an ongoingbasis. In PPC, the sustained evolution in time of the populationactivity lasting throughout movement duration probably reflectsspecification and control of the kinematics of the new trajectory,regardless of when the second target is presented.

5. Time for visuomotor transformations and coexistence ofneural signals concerning different motor plans

An interesting aspect of the online control of hand movementtrajectory refers to when, in the cerebral cortex, the visual signalabout target location influences neural activity. This can be studiedby the correlation, in the same neuron, between its activity insingle-step reaches with that during double-step reaches to thesame first target. In PPC, this correlation is initially high, thenabruptly decrease at about 150 ms after the second target presen-tation. The same 150 ms interval is observed regardless of whetherthe second target is presented during reaction time or at the begin-ning of hand movement time. Therefore, this interval may be con-sidered as the time necessary for the visuomotor transformationunderlying reaches to visual targets.



527

528

529

530

531

532

533

534

535

536

537

538

539

540

541

542

543

544

545

546

547

548

549

550

551

552

Fig. 4. Predicting cell activity in PMd, M1 and PPC during corrected reaches from cell activity observed during direct reaches. Comparison between cell activity during double-step trials (black spike-density functions, SDFs) from the center (C) to one peripheral target (A) that jumps in the opposite direction (B), and cell activity ‘‘reconstructed’’ bycombining, tip-to-tail, the two spike density functions (gray curves) associated with single-step trials from C to A and from C to B (see bottom panel) of the same cell. Thetarget was switched during hand reaction- or at the onset of movement-time. In each SDF, the vertical dashed lines represent the time at which cell activity associated to thefirst reach movement (from C to target A) is truncated and replaced by the activity associated to the second hand movement from C to B. T, target presentation; TJ, targetjump; M, movement onset; HS, switch of hand movement direction.



30 September 2014

A further and related question refers the potential coexis-tence of neural signals concerning the planning of a new trajec-tory during performance of a movement specified by theprevious motor plan. Such coexistence is possible if the late partof hRT2, which encodes the future hand movement, overlaps intime with the hand movement specified by the previous motorcommand; this is with the assumption that the first 150 ms ofhRT2 are devoted to the visuomotor transformation, as dis-cussed above. One observation supporting this statement is thatin PPC and M1, when the second target is presented duringhRT1, the correlation in firing frequencies between single anddouble-step reaches drops off at the onset of hand movement(Fig. 6A), even though the hand trajectories in both conditions


are still identical at that particular point in time. By contrast,when the second target is presented at the onset of hMT, thiscorrelation begins to decrease 150 ms after the initiation of handmovement. Interestingly, the same phenomenon is observed inPMd, where however the decrease in correlation occurs earlierin time during the trial. This would suggest that PMd is the firstcortical area to be influenced by the coexistence of old and newmotor intentions, and therefore to signal the need to changehand movement trajectory.

In other types of tasks where the monkey is given more time todecide its next action, the coexistence of neural signals for reachingmovements to two distinct targets can last longer (Cisek & Kalaska,2005, 2010). In the experimental setup of Cisek and Kalaska, ani-



553

554

555

556

557

558

559

560

561

562

563

564

565

566

567

568

569

570

571

572

573

574

575

576

577

578

579

580

581

582

583

584

585

586

587

588

589

590

591

592

593

594

595

596

597

Fig. 5. Temporal evolution of the population activity of reach-related cells in theparieto-frontal system. Comparisons of population SDF (pop-SDF) of the neuralactivity recorded during direct reaches (faint colors) versus corrected reaches(bright colors), when the target jumped during RT. For the on-line corrections thepop-SDF were obtained averaging all the single-cell activities for hand movementsfirst directed toward the target opposite to the preferred direction (anti-PD) andthen to the PD. This activity is compared to direct reaches toward the first target(anti-PD) to depict the time at which the two population activities diverge. Thetriangles indicate the times of target presentation (T) for both direct and correctedreaches and target jump (TJ) during double-step reaches, while the horizontal barsindicate the mean duration of movement time, for direct reaches (faint color) andfor corrected ones (bright color). The time scale is aligned to the onset of handmovement (vertical black line; 0 ms). The vertical dashed lines indicate the time atwhich neural activity in corrected reaches significantly diverges from that of thedirect reaches, while the colored vertical solid line indicate the mean time of changeof hand trajectory. (Reproduced with modifications from Archambault, Ferrari-Toniolo, & Battaglia-Mayer, 2011.)

Fig. 6. (A) Comparing neural activity during direct and corrected reaches. Evolutionin time of the correlation coefficient (R) between neural activity recorded in PMd(red), M1 (green) and PPC (blue) during direct and corrected reaches with targetjump occurring during hRT1. Data are shown within a time window spanning from400 ms before to 400 ms after h MT onset (0 ms). (B) Schematic representation oftasks (single-step and double-step with target jump during hRT1) and mainbehavioral events, whose mean values of time of occurrence are reported on thetime axis shown in A (dashed lines). Thick horizontal bars represent the relativetiming of the central and peripheral target presentation in the different tasks (gray:single-step, black: double-step). Curves represent the schematic temporal evolutionof hand position (gray: single-step, black: double-step). (For interpretation of thereferences to color in this figure legend, the reader is referred to the web version ofthis article.)



30 September 2014

mals are first presented with two potential reach targets; then thetargets disappear for an interstimulus interval of 1–1.5 s. Finally, anon-spatial cue is provided signalling which target should bereached. During the interstimulus interval, the authors recordedin PMd coexisting, directional activity related to the two potentialmovements. Therefore, for the coexistence of neural activity to bedetected in a serial task such as the single-step/double-step para-digm, the presentation of the second target should occur within agiven time interval. This insures that the final part of the reactiontime to the second target (hRT2), which carries information aboutthe new target direction, overlaps with the hand movement tothe first target (Fig. 6B). And in parallel tasks where the choicebetween two targets has to be maintained in memory, then coex-isting signals are observed until the monkey performs the reachingmovement.


6. The causal relationship between cell activity in parietalcortex and online correction: reversible inactivation of superiorparietal areas in monkeys

To assess the existence of causal relationships between cell activ-ity and the ability to adjust hand movement trajectory, a reversibleinactivation experiment has been performed in macaque monkeys(Battaglia-Mayer et al., 2013). This involved injecting the GABA-Aagonist muscimol into the parietal areas, where cell activity duringsingle- and double-step reaches had been fully characterized.

Muscimol was injected either unilaterally in the left or rightsuperior parietal lobule (SPL, area 5, areas PE/PEc) or bilaterally.It was observed that unilateral inactivation of the SPL only had aslight effect on the hand reaching performance, by increasing thevariability of the trajectory but with no significant changes inhRT or hMT, as compared to the control conditions (no injection)or to saline injection. Bilateral deactivation of SPL (Fig. 7A and B)led to important changes in trajectory variability, as well as signif-icant increases in hRT1, hRT2 and hMT for both single- and double-step trials, as compared to control conditions. Eye behavior waslikewise affected by bilateral muscimol injection, with significantlyhigher eRT1 and eRT2 compared to control conditions (Battaglia-Mayer et al., 2013). The eye impairment was, at least in part,responsible for the delayed hand movement correction. Theseresults point to a direct involvement of the PPC in the online con-trol of both hand and eye movements with a switch in target. It ispossible that bilateral PPC deactivation was necessary to observe adeficit in online control due to the extensive training that the ani-mals have received in this experimental paradigm. Indeed, follow-ing task-specific training, motor maps associated with a movementor body part in M1 expand to neighboring areas, a phenomenon



598

599

600

601

602

603

604

605

606

607

608

609

610

611

612

613

614

615

616

617 Q2

618

619

620

621

622

623

624

625

626

627

628

629

630

631

632

633

634

635

Fig. 7. Monkey model of optic ataxia. Consequences of inactivation of superior parietal areas PE/PEc on trajectory length (A) and dispersion (B). (A) Replications of individualhand trajectories (thin curves), and their mean (thick curve) from a central position (C) during corrected reaches with the right arm, after target displacement (from target Ato target B) at the onset of hand movement before (green, ‘‘control’’ condition) and after bilateral inactivation (red). (B) Modification in the dispersion of trajectories undercontrol conditions (control and saline injection), and unilateral and bilateral parietal inactivation. Cumulative frequency distributions of the correlation coefficients obtainedby comparing the hand trajectories under different conditions to those recorded before muscimol injection. The latter are compared to their own mean (green). Thedistribution referring to bilateral inactivation (red) is significantly (K–S test, p < 0.001) different from control conditions (green), and after unilateral inactivation or salineinjection, that showed little or no effects relative to controls. (Reproduced with modifications from Battaglia-Mayer et al., 2013.) (For interpretation of the references to colorin this figure legend, the reader is referred to the web version of this article.)



30 September 2014

which is thought to occur in other brain areas as well (Dancause &Nudo, 2011; Nudo et al., 1996). Thus, more extensive cortical deac-tivation may be necessary in well-trained animals, in order toobserve an impairment during double-step reaching.

636

637

638

639

640

641

642

643

644

645

646

Fig. 8. The cortical network underlying the visual control of hand movement.Figurine of the monkey brain showing the cortical areas where cell activity has been(so far) shown to be modulated during online correction of hand movement (PE/PEc, M1, PMd, PMv; red ovals) with their known cortico-cortical connections,embedded with the more general connectivity scheme depicting the main sourcesof visual input to the PMd and M1 for the composition and control of visually-guided reach movements. This input stems, at least from V6A, 7m, and MIP. AS:arcuate sulcus; CS: central sulcus; IPS: inferior parietal sulcus. (For interpretation ofthe references to color in this figure legend, the reader is referred to the web versionof this article.)

7. The sources of the visual input for composition and onlinecorrection of hand movement trajectory in the parieto-frontalsystem

The ability to control hand movement and to update its trajec-tory at any moment indicates that visual information about targetlocation, which is crucial for the composition and control of reach-ing movements, has a continuous access to the frontal lobe motorareas that ultimately command arm movement. The source of thevisual information for reach control was first described byCaminiti, Ferraina, and Johnson (1996), Johnson et al. (1996) andJohnson, Ferraina, and Caminiti (1993). The regions (areas PE,PEc) of the superior parietal lobule and of the frontal cortex(PMd/area 6; M1/area 4), where cell activity related to online con-trol has been revealed, are linked by direct (Innocenti, Vercelli, &Caminiti, 2014; Johnson et al., 1996; Marconi et al., 2001; Matelliet al., 1998), and indirect, via PMd (Bakola et al., 2010; Innocentiet al., 2013; Johnson et al., 1996; Matelli et al., 1998) cortico-corti-cal connections (Fig. 8). In area PEc, neurons are sensitive, amongothers things, to reach-related signals such as hand and/or eyeposition and movement direction (Battaglia-Mayer et al., 2001),to the retinal position and direction of motion of visual stimuli(Battaglia-Mayer et al., 2000, 2001) and to reach distance(Ferraina et al., 2009) and combine all this information within theirglobal tuning field (Battaglia-Mayer et al., 2001). Local parietal pro-jections to PEc stem from a distributed network, including PEa,MIP, PEci, and, to a lesser extent, 7m, V6A (part of area 19), 7a,MST (Marconi et al., 2001). Therefore this posterior part of SPL,including areas V6A, PEc, and 7m, which is located just anteriorto the parieto-occipital sulcus and on the medial wall, can be con-sidered as the source of visual input to the frontal premotor andmotor areas for control and update of visual reaching.

647

648

649

650

651

8. Neuroimaging, cortical lesions and stimulation studies

Using positron-emission tomography (PET), Desmurget et al.(2001) contrasted double-step to singles-step hand reaching


movements, with the shift in target occurring during the ocularsaccade to the first target (meaning that participants were not con-sciously aware of their own corrected movement). Three brainareas were highlighted as being more active during double- versussingle-step trials, namely those of the intraparietal sulcus (IPS), theM1 hand area and the right anterior parasagittal cerebellar cortex,which is linked to arm movements. This finding is coincident withmonkey neurophysiological research, where the same corticalareas (PPC and M1) were found to have cells whose activity waslinked to hand trajectory update during single- and double-stepreaching tasks (Archambault, Caminiti, & Battaglia-Mayer, 2009;Archambault, Ferrari-Toniolo, & Battaglia-Mayer, 2011; Dickey,Amit, & Hatsopoulos, 2013).

The relationship between brain activity and behavior can alsobe probed through lesion studies. These lesions can occur inhumans as part of neurological conditions (i.e., stroke). In addition,



652

653

654

655

656

657

658

659

660

661

662

663

664

665

666

667

668

669

670

671

672

673

674

675

676

677

678

679

680

681

682

683

684

685

686

687

688

689

690

691

692

693

694

695

696

697

698

699

700

701

702

703

704

705

706

707

708

709

710

711

712

713

714

715

716

717

718

719

720

721

722

723

724

725

726

727

728

729

730

731

732

733

734

735

736

737

738

739

740

741

742

743

744

745

746

747

748

749

750

751

752

753

754

755

756

757

758

759

760

761

762

763

764

765

766

767

768

769

770

771

772

773

774

775

776

777

778



30 September 2014

a technique such as transcranial magnetic stimulation (TMS) canbriefly disrupt activity in a constrained cortical region. In monkeys,muscimol (a GABA-A agonist) can be injected in a patch of cortex inorder to reversibly disrupt activity for a period of few hours. Allthree types of studies have been reported in the literature, in thecontext of online correction of hand movements.

Stroke often affect multiple cortical and subcortical areas andthose constrained to the PPC are very rare. Pisella et al. (2000)studied one patient with bilateral PPC lesions causing optic ataxiaand found that, compared to healthy controls, this patient was ableto reach for fixed (foveated) targets without any apparent deficit(single-step trials). However, when the target was displaced in adouble-step paradigm, the patient was unable to produce fast, cor-rected hand movements, performing only slow corrections. TheLyon group interpreted these observations as more compatiblewith a specific role of PPC in processing online (‘automatic’) correc-tions in response to a location perturbation, rather being responsi-ble of more general visuomotor functions. In a second study, thesame authors found similar results when the patient was requiredto reach and grasp a cylinder; if the cylinder’s position unexpect-edly shifted during movement, then the patient first completed areaching movement to the original target’s location, before per-forming a second movement to the new location (Grea et al.,2002). Moreover, Buiatti, Skrap, and Shallice (2013) in a large ana-tomical based group study found that parietal patients with opticataxia are affected during direct unperturbed reaches and onlinecorrection of movement trajectory.

Various groups also studied the effects of TMS during eitherhand reaching or grasping, with a sudden shift in the target’s loca-tion or size. It was found that TMS applied during hRT2 over the IPScould disrupt double-step movements by either making subjectsfail to correct their movements for small shifts in target position(Desmurget et al., 1999), or could lengthen total movement timewith larger target shifts (Reichenbach et al., 2011), when comparedto double-step trials without TMS. A similar phenomenon wasobserved when subjects had to grasp a target that could suddenlychange in size while remaining at the same location; TMS appliedto the IPS, but not M1, was then seen to lengthen the total move-ment time for corrected movements (Glover, Miall, & Rushworth,2005; Tunik, Frey, & Grafton, 2005). In later studies involvingTMS during reach and grasp movements, it was found that stimu-lation to the anterior part of the IPS, and not the caudal, and to thecontralateral hemisphere, and not the ipsilateral, disrupted adjust-ments to a sudden change in size (Rice et al., 2007; Rice, Tunik, &Grafton, 2006). Note however that other researchers, using thesame method as Desmurget et al., failed to measure any changein endpoint accuracy or movement time during double-step trials,with TMS applied to the IPS (Johnson & Haggard, 2005). It is possi-ble that the stimulation in that experiment was less accurate, aspositioning of the TMS probe was not guided by an MRI image. Afurther evidence that points to the role of PPC in the adjustmentsof motor commands is provided by an experiment in which PPCwas stimulated with TMS, while subjects learned to make reachingmovement in a velocity-depended force-field (Della-Maggioreet al., 2004). The control subject, who received TMS pulses overthe occipital cortex, learned to cope with the new dynamics, asshown by their movement curvature gradually decreasing to base-line level. On the contrary, reach direction errors remained signif-icantly larger in the group where PPC was perturbed, thusdemonstrating that the alteration of parietal activity has a signifi-cant impact on the ability to generate corrective movements.

Another cortical area involved in on-line correction of handmovements is the dorsal premotor cortex (PMd), but only whentrajectory correction depends on visual information concerningvision of the hand during a visuo-motor adaptation task (Lee &van Donkelaar, 2006). In absence of such visual feedback, TMS


perturbation of PMd does not affect on-line adjustments, as it doeswhen vision of the hand is available, suggestive of an early signal-ling for on-line correction occurring in PMd. Patients with PMdlesion show increased hand movement-time when the targetmoves in space (Buiatti, Skrap, & Shallice, 2013), but do not showthe reach endpoint errors typical of parietal patients, in otherwords they pay a cost for reach timing but not for reach accuracy.

9. A ‘‘positive image’’ of optic ataxia of parietal patients

The failure of patients suffering from optic ataxia to make fast,online corrections of their hand movement trajectories might bedependent on the loss cells in the PPC, whose activity carries infor-mation concerning corrections of hand movement direction. Theactivity of cells in the SPL and at the parieto-occipital junctioncombine the position and direction of coordinated eye–handmovements, as well as the position and direction of motion ofvisual stimuli within global tuning fields (Battaglia-Mayer &Caminiti, 2002). Their orientation cover in a uniform fashion theeye and hand action space. Thus, these cells encode all the direc-tional signals necessary to compose and update hand movementtrajectory in a congruent fashion. The collapse of this combinato-rial mechanisms, in other words the visuo-motor transformationunderlying hand movement, will impair both the reaching move-ment and its modification, as seen in parietal patients sufferingfrom optic ataxia. Therefore, the results of the neurophysiologicalstudies of PPC in monkeys provide a ‘‘positive image’’ of some cru-cial features observed in optic ataxia patients (see Battaglia-Mayeret al., 2014 for a discussion).

10. Theories of online motor control

The online correction of hand movements in humans has beenexamined from the point of view of various theories of motor con-trol. Indeed, researchers have attempted to explain the specificshape of hand movement trajectories, with and without a changein target location, and from these to draw inferences about howmovement is planned and executed by the CNS. For example, theminimum jerk model states that natural point-to-point move-ments follow a trajectory, which minimizes the magnitude of ‘jerk’,the time derivative of acceleration (Flash & Hogan, 1985; Viviani &Flash, 1995). Minimization of jerk was observed in a variety ofmovements (point-to-point, through an intermediate point,around obstacles, etc.) and was interpreted as a goal by the CNSto produce the smoothest possible movement.

In a paper by Flash and Henis (1991), jerk minimization wasapplied to movement trajectories involving a shift in target loca-tion. They introduced two hypotheses of movement execution:the ‘‘replacement hypothesis’’ assumes that, after a switch in targetlocation, control of hand movement toward the first target isaborted and replaced to a plan driving the hand to the second tar-get. As an alternative, the authors also introduced a ‘‘superpositionhypothesis’’, whereas after a target switch, the motor plan towardthe first target continues as initially intended, while a second planis added vectorially; that second plan would be one moving thehand from the original target location to its displaced position.Simulations of movement involving minimum jerk optimizationsupported the superposition over the replacement hypothesis.

Another theoretical approach, the equilibrium point hypothesis(k model), posits that the CNS does not program exact muscleforces and joint torques, but would control a referent trajectoryfor reaching movements (Feldman & Levin, 1995). The actualmovement is then the result of the dynamic interaction between,on one hand, the referent trajectory, and, on the other, peripheralfeedback (reflexes) as well as biomechanical properties of the



779

780

781

782

783

784

785

786

787

788

789

790

791

792

793

794

795

796

797

798

799

800

801

802

803

804

805

806

807

808

809

810

811

812

813

814

815

816

817

818

819

820

821

822

823

824

825

826

827

828

829

830

831

832

833

834

835

836

837

838

839

840

841

842

843

844

845

846

847

848

849

850

851

852

853

854

855

856

857

858

859

860

861

862

863

864

865

866

867

868

869

870

871

872



30 September 2014

muscles and joints (stiffness, damping, etc.). Flanagan, Ostry, andFeldman (1993) were able to simulate movements involving a shiftin target using the k model together with the superpositionhypothesis, and obtained an excellent match between the simu-lated and observed trajectories.

Thus, both the superposition and replacement hypotheses maybe valid, at least from the two theoretical models described above,but what links can be made with existing neurophysiological data?Simultaneous encoding of motor plans has been observed in PMdcells, in tasks when monkeys are presented with two possible reachtargets and wait for a non-spatial go-signal. In this condition, differ-ent populations of cells are active simultaneously, reflecting thetwo possible future movements. At first glance, this phenomenoncould be linked to Flahs and Hennis’ superposition hypothesis.However, the superposition and replacement hypotheses are sup-posed to describe movement execution, not planning; and in Cisekand Kalaska’s work, the simultaneous encoding of two movementdirections coalesces to a single encoding just before the onset ofhand movement. In other words, when the monkey decides to actu-ally move to one of the two targets, the activity related to the plan-ning of movement to the other target completely disappears.

Dickey et al. have looked specifically at neural activity duringthe transition from the first to the second movement during dou-ble-step tasks. They found neurons in M1 and PMd where theactivity changed abruptly from one direction to the next, and a lar-ger group of neurons where the change in activity was smooth.They likened the abrupt transition with the replacement hypothe-sis, and the smooth one with the superposition hypothesis. We donot totally agree with this interpretation, however, as the superpo-sition hypothesis of Flash and Hennis implies simultaneous, paral-lel commands throughout movement execution. Thus, neuralactivity during double-step tasks appears to be more consistentwith the replacement hypothesis of Flash and Hennis and Flanaganet al, with a brief transition phase following the switch from thefirst to the second target, where both targets may be encoded.

873

874

875

876

877

878

879

880

881

882

883

884

885

886

887

888

889

890

891

11. Conclusions

Taken together, the neurophysiological, neuroimaging, lesionand inactivation studies reviewed in the previous sectionsoverwhelmingly point to an important role of the PPC in the onlinecorrection of hand reaching movements, in both human and non-human primates (Battaglia-Mayer et al., 2014; Gaveau et al.,2014). This is consistent with the well-known role of the PPC inthe coordination of visually guided hand movements, which isexpressed both during the planning and the execution of move-ment. Anatomically, the PPC receives somatosensory and visualinformation, and has bidirectional connections with the pre-motorcortex, which in turns projects to M1. The more anterior part of SPLalso projects directly to M1 (Johnson et al., 1996; Marconi et al.,2001). Thus, it is ideally positioned to estimate the state of the handmovement and task requirements, and convey this information toM1 and PMd (Desmurget et al., 2001).

892

893

894

895

896

897

898

899

900

901

902

903

11.1. Comparing human and monkey experiments

Despite the paucity of studies allowing a rigorous comparisonbetween results obtained in humans and monkey, an attemptcan be made on the basis of the available data. On one hand, theimpairments observed after bilateral inactivation through musci-mol injections resemble that observed in patient I.G., with bilateralparietal lesions: a deficit in adjusting online the hand trajectory toa new target location. On the other hand, similar deficits in onlinecontrol of hand movements are observed after unilateral deactiva-tion in humans, through TMS. This apparent discrepancy can be


attributed to the difficulty in directly comparing the TMS and mus-cimol deactivation, due to the lack of evidence about the extensionof the lesioned/inactivated regions, in either case. In humans, alarge study group has shown difficulties in online control after uni-lateral parietal lesion in optic ataxia patients (Buiatti, Skrap, &Shallice, 2013). Therefore, it can be argued that the lateralizationof spatial functions to the right PPC in humans, as well the deacti-vation injection area through muscimol in monkeys, both contrib-ute to this apparent discrepancy.

11.2. Understanding the neurophysiology of online control

A baffling result of cell recording studies is that in the parietalcortex no cells have been observed that only fire during online cor-rection, or that would encode corrective movements only, despiteits central role in this regard, as evidenced by lesion and inactiva-tion studies, in both humans and monkeys. In addition, when look-ing at the relative timing of different cortical areas, PMd activity is,on average, earlier in time than M1 and PPC activity. How can thisbe reconciled with the fact that PPC plays a crucial role in onlinecorrection, for both monkeys and humans? Unfortunately, thereare no studies on human motor cortex (area 4, M1) that showthe consequences of lesions or inactivation on corrective move-ments. The few available in the premotor (area 6, PMd) cortex con-firm the role of this area in the selection of future movements. InSchluter et al.’s (1998), TMS applied over the premotor areas dis-rupt an early stage of movement selection, while motor cortexstimulation disrupts the movements at a later stage of execution.From a different perspective, Lee and van Donkelaar (2006)showed that TMS perturbation of PMd affects online adjustmentsonly when vision of the hand is available. Buiatti, Skrap, andShallice (2013) have shown that patients with premotor lesionsare slower in trajectory correction, but not in accuracy. Thus thepremotor cortex might provide the higher-order command signalto change a motor plan and movement direction. The PPC activitycan be seen to change at all times during the task, and importantly,during double-step movements, thus providing a continuous esti-mate of limb kinematics, as well as of the direction of dynamicforce, as suggested by a recent case report (Ferrari-Toniolo et al.,2014) of a parietal patient displaying a strong impairment in thespecification of the direction of the force necessary to move a cur-sor to visual targets in different directions in an isometric task.Interestingly, the specification of force magnitude was not affectedin this patient. As far as motor cortex is concerned, the strong rela-tionship of cell activity with hand kinematics and the fact that thisactivity leads movement both suggest a pivotal role in the preciseimplementation of the movement trajectory on an ongoing basis.In fact, motor cortex can be directly influenced by cortico-corticalconnections from PMd as well as via PPC, as suggested by the rel-ative times in different structures at which the activity in correctedreaches diverges from the uncorrected ones. Moreover, the privi-leged access of motor cortex to the command system of the spinalcord and the relationships between cell activity in this area anddynamic force (Georgopoulos et al., 1992) suggest a potential rolein encoding the force necessary to brake the hand movement to thefirst target, and then accelerate it toward the final destination.

No study has reported any evidence of selected population ofcells, in either M1, PMd or PPC, that are modulated only duringmovement correction and not during simple, unperturbed reaches.Therefore, the planning and correction of reaches can only reside ina graded and time-varying utilization of the kinematic variablesthat are distributed throughout the cortical assemblies that, in dif-ferent areas, are devoted to movement production and online con-trol. The fact that the same time-varying muscle synergies controlboth direct and corrected reaches also supports this view (seeBattaglia-Mayer et al., 2014).



904

905

906

907

908

909

910

911

912

913

914

915

916

917

918

919

920

921

922 Q3

923

924 Q4

925 Q5

926

927

928929930931932933934935936937938939940941942943944945946947948949950951952953954955956957958959960961962963964965966967968969970971972973974

9759769779789799809819829839849859869879889899909919929939949959969979989991000100110021003100410051006100710081009101010111012101310141015101610171018101910201021102210231024102510261027102810291030103110321033103410351036103710381039104010411042104310441045104610471048104910501051105210531054105510561057105810591060



30 September 2014

It has been suggested that online correction of hand movementsis based on a comparison of the ongoing movement with an effer-ence copy, or the nervous system’s internal representation ofmovement execution (Davidson & Wolpert, 2005; Desmurget &Grafton, 2000). Such a comparison requires both feedforward andfeedback processes, which is similar to what has been observedin the PPC during online correction movements in monkeys(Archambault, Caminiti, & Battaglia-Mayer, 2009). In other words,while cells in the PPC may, on average, fire later with respect tohand movements than cells in PMd and M1, their activity seemsto be dependent on sensory feedback, more than in these frontalareas. Based on the idea of the efference copy, it can be argued thatit is an optimal blending of cells combining feedforward and feed-back signals that is crucial for the online correction of hand move-ments in PPC in order to contribute to online correction.

12. Uncited references

Izawa and Shadmehr (2008), Shadmehr and Krakauer (2008),Vaziri, Diedrichsen, and Shadmehr (2006) and Wolpert,Ghahramani, and Jordan (1995).

Acknowledgments

This work was supported by the MIUR of Italy (prot.2010MEFNF7_004 to RC) and by the Quebec Research Funds forHealthcare (FRQS).

References

Archambault, P. S., Caminiti, R., & Battaglia-Mayer, A. (2009). Cortical mechanismsfor online control of hand movement trajectory: The role of the posteriorparietal cortex. Cerebral Cortex, 19(12), 2848–2864.

Archambault, P. S., Ferrari-Toniolo, S., & Battaglia-Mayer, A. (2011). Online controlof hand trajectory and evolution of motor intention in the parietofrontalsystem. Journal of Neuroscience, 31(2), 742–752.

Bakola, S., Gamberini, M., Passarelli, L., Fattori, P., & Galletti, C. (2010). Corticalconnections of parietal field PEc in the macaque: Linking vision and somaticsensation for the control of limb action. Cerebral Cortex, 20(11), 2592–2604.

Battaglia-Mayer, A., Buiatti, T., Caminiti, R., Ferraina, S., Lacquaniti, F., & Shallice, T.(2014). Correction and suppression of reaching movements in the cerebralcortex: Physiological and neuropsychological aspects. Neuroscience &Biobehavioral Reviews, 42, 232–251.

Battaglia-Mayer, A., & Caminiti, R. (2002). Optic ataxia as a result of the breakdownof the global tuning fields of parietal neurones. Brain, 125(Pt 2), 225–237.

Battaglia-Mayer, A., Caminiti, R., Lacquaniti, F., & Zago, M. (2003). Multiple levels ofrepresentation of reaching in the parieto-frontal network. Cerebral Cortex,13(10), 1009–1022.

Battaglia-Mayer, A., Ferraina, S., Genovesio, A., Marconi, B., Squatrito, S., Molinari,M., et al. (2001). Eye–hand coordination during reaching. II. An analysis of therelationships between visuomanual signals in parietal cortex and parieto-frontal association projections. Cerebral Cortex, 11(6), 528–544.

Battaglia-Mayer, A., Ferraina, S., Mitsuda, T., Marconi, B., Genovesio, A., Onorati, P.,et al. (2000). Early coding of reaching in the parietooccipital cortex. Journal ofNeurophysiology, 83(4), 2374–2391.

Battaglia-Mayer, A., Ferrari-Toniolo, S., Visco-Comandini, F., Archambault, P. S.,Saberi-Moghadam, S., & Caminiti, R. (2013). Impairment of online control ofhand and eye movements in a monkey model of optic ataxia. Cerebral Cortex,23(11), 2644–2656.

Brenner, E., & Smeets, J. B. (1997). Fast responses of the human hand to changes intarget position. Journal of Motor Behavior, 29(4), 297–310.

Briere, J., & Proteau, L. (2011). Automatic movement error detection and correctionprocesses in reaching movements. Experimental Brain Research, 208(1), 39–50.

Buiatti, T., Skrap, M., & Shallice, T. (2013). Reaching a moveable visual target:Dissociations in brain tumour patients. Brain and Cognition, 82(1), 6–17.

Caminiti, R., Chafee, M. V., Battaglia-Mayer, A., Averbeck, B. B., Crowe, D. A., &Georgopoulos, A. P. (2010). Understanding the parietal lobe syndrome from aneurophysiological and evolutionary perspective. European Journal ofNeuroscience, 31(12), 2320–2340.

Caminiti, R., Ferraina, S., & Johnson, P. B. (1996). The sources of visual information tothe primate frontal lobe: A novel role for the superior parietal lobule. CerebralCortex, 6(3), 319–328.

Cisek, P., & Kalaska, J. F. (2005). Neural correlates of reaching decisions in dorsalpremotor cortex: Specification of multiple direction choices and final selectionof action. Neuron, 45(5), 801–814.

Cisek, P., & Kalaska, J. F. (2010). Neural mechanisms for interacting with a world fullof action choices. Annual Review of Neuroscience, 33, 269–298.


Dancause, N., & Nudo, R. J. (2011). Shaping plasticity to enhance recovery afterinjury. Progress in Brain Research, 192, 273–295.

d’Avella, A., Portone, A., Fernandez, L., & Lacquaniti, F. (2006). Control of fast-reaching movements by muscle synergy combinations. Journal of Neuroscience,26(30), 7791–7810.

d’Avella, A., Portone, A., & Lacquaniti, F. (2011). Superposition and modulation ofmuscle synergies for reaching in response to a change in target location. Journalof Neurophysiology, 106(6), 2796–2812.

Davidson, P. R., & Wolpert, D. M. (2005). Widespread access to predictive models inthe motor system: A short review. Journal of Neural Engineering, 2(3), S313–S319.

Della-Maggiore, V., Malfait, N., Ostry, D. J., & Paus, T. (2004). Stimulation of theposterior parietal cortex interferes with arm trajectory adjustments during thelearning of new dynamics. Journal of Neuroscience, 24(44), 9971–9976.

Desmurget, M., Epstein, C. M., Turner, R. S., Prablanc, C., Alexander, G. E., & Grafton,S. T. (1999). Role of the posterior parietal cortex in updating reachingmovements to a visual target. Nature Neuroscience, 2(6), 563–567.

Desmurget, M., & Grafton, S. (2000). Forward modeling allows feedback control forfast reaching movements. Trends in Cognitive Sciences, 4(11), 423–431.

Desmurget, M., Grea, H., Grethe, J. S., Prablanc, C., Alexander, G. E., & Grafton, S. T.(2001). Functional anatomy of nonvisual feedback loops during reaching: Apositron emission tomography study. Journal of Neuroscience, 21(8), 2919–2928.

Dickey, A. S., Amit, Y., & Hatsopoulos, N. G. (2013). Heterogeneous neural coding ofcorrective movements in motor cortex. Frontiers in Neural Circuits, 7, 51.

Fautrelle, L., Prablanc, C., Berret, B., Ballay, Y., & Bonnetblanc, F. (2010). Pointing todouble-step visual stimuli from a standing position: Very short latency(express) corrections are observed in upper and lower limbs and may notrequire cortical involvement. Neuroscience, 169(2), 697–705.

Feldman, A. G., & Levin, M. F. (1995). The origin and use of positional frames ofreference in motor control. Behavioral and Brain Sciences, 18, 723–806.

Ferraina, S., Battaglia-Mayer, A., Genovesio, A., Archambault, P., & Caminiti, R. (2009).Parietal encoding of action in depth. Neuropsychologia, 47(6), 1409–1420.

Ferrari-Toniolo, S., Papazachariadis, O., Visco-Comandini, F., Salvati, M., D’elia, A.,Caminiti, R., et al. (2014). A visuomotor disorder in absence of movement: DoesOptic Ataxia generalize to learned isometric hand action? Neuropsychologia, 63,59–71.

Flanagan, J. R., Ostry, D. J., & Feldman, A. G. (1993). Control of trajectorymodifications in target-directed reaching. Journal of Motor Behavior, 25(3),140–152.

Flash, T., & Henis, E. (1991). Arm trajectory modifications during reaching towardsvisual targets. Journal of Cognitive Neuroscience, 3(3), 220–230.

Flash, T., & Hogan, N. (1985). The coordination of arm movements: Anexperimentally confirmed mathematical model. Journal of Neuroscience, 5(7),1688–1703.

Franklin, D. W., & Wolpert, D. M. (2008). Specificity of reflex adaptation for task-relevant variability. Journal of Neuroscience, 28(52), 14165–14175.

Gaveau, V., Pisella, L., Priot, A. E., Fukui, T., Rossetti, Y., Pelisson, D., et al. (2014).Automatic online control of motor adjustments in reaching and grasping.Neuropsychologia, 55, 25–40.

Georgopoulos, A. P., Ashe, J., Smyrnis, N., & Taira, M. (1992). The motor cortex andthe coding of force. Science, 256, 1692–1695.

Georgopoulos, A. P., Kalaska, J. F., Caminiti, R., & Massey, J. T. (1983). Interruption ofmotor cortical discharge subserving aimed arm movements. Experimental BrainResearch, 49(3), 327–340.

Georgopoulos, A. P., Kalaska, J. F., & Massey, J. T. (1981). Spatial trajectories andreaction times of aimed movements: Effects of practice, uncertainty, andchange in target location. Journal of Neurophysiology, 46(4), 725–743.

Gielen, C. C., van den Heuvel, P. J., & Denier van der Gon, J. J. (1984). Modification ofmuscle activation patterns during fast goal-directed arm movements. Journal ofMotor Behavior, 16(1), 2–19.

Glover, S., Miall, R. C., & Rushworth, M. F. (2005). Parietal rTMS disrupts theinitiation but not the execution of on-line adjustments to a perturbation ofobject size. Journal of Cognitive Neuroscience, 17(1), 124–136.

Goodale, M. A., Pelisson, D., & Prablanc, C. (1986). Large adjustments in visuallyguided reaching do not depend on vision of the hand or perception of targetdisplacement. Nature, 320(6064), 748–750.

Grea, H., Pisella, L., Rossetti, Y., Desmurget, M., Tilikete, C., Grafton, S., et al. (2002). Alesion of the posterior parietal cortex disrupts on-line adjustments duringaiming movements. Neuropsychologia, 40(13), 2471–2480.

Gritsenko, V., Yakovenko, S., & Kalaska, J. F. (2009). Integration of predictivefeedforward and sensory feedback signals for online control of visually guidedmovement. Journal of Neurophysiology, 102(2), 914–930.

Innocenti, G. M., Vercelli, A., & Caminiti, R. (2014). The diameter of cortical axonsdepends both on the area of origin and target. Cerebral Cortex, 24, 2178–2188.

Izawa, J., & Shadmehr, R. (2008). On-line processing of uncertain information invisuomotor control. Journal of Neuroscience, 28(44), 11360–11368.

Johnson, P. B., Ferraina, S., Bianchi, L., & Caminiti, R. (1996). Cortical networks forvisual reaching: Physiological and anatomical organization of frontal andparietal lobe arm regions. Cerebral Cortex, 6(2), 102–119.

Johnson, P. B., Ferraina, S., & Caminiti, R. (1993). Cortical networks for visualreaching. Experimental Brain Research, 97(2), 361–365.

Johnson, H., & Haggard, P. (2005). Motor awareness without perceptual awareness.Neuropsychologia, 43(2), 227–237.

Johnson, H., Van Beers, R. J., & Haggard, P. (2002). Action and awareness in pointingtasks. Experimental Brain Research, 146(4), 451–459.

Kadota, K., & Gomi, H. (2010). Implicit visuomotor processing for quick onlinereactions is robust against aging. Journal of Neuroscience, 30(1), 205–209.



10611062106310641065106610671068106910701071107210731074107510761077107810791080108110821083108410851086 Q61087108810891090109110921093109410951096109710981099110011011102

110311041105110611071108110911101111111211131114111511161117111811191120112111221123112411251126112711281129113011311132113311341135113611371138113911401141114211431144



30 September 2014

Knill, D. C., Bondada, A., & Chhabra, M. (2011). Flexible, task-dependent use ofsensory feedback to control hand movements. Journal of Neuroscience, 31(4),1219–1237.

Lee, J. H., & van Donkelaar, P. (2006). The human dorsal premotor cortex generateson-line error corrections during sensorimotor adaptation. Journal ofNeuroscience, 26(12), 3330–3334.

Leonard, J. A., Gritsenko, V., Ouckama, R., & Stapley, P. J. (2011). Posturaladjustments for online corrections of arm movements in standing humans.Journal of Neurophysiology, 105(5), 2375–2388.

Liu, D., & Todorov, E. (2007). Evidence for the flexible sensorimotor strategiespredicted by optimal feedback control. Journal of Neuroscience, 27(35),9354–9368.

Marconi, B., Genovesio, A., Battaglia-Mayer, A., Ferraina, S., Squatrito, S., Molinari,M., et al. (2001). Eye–hand coordination during reaching. I. Anatomicalrelationships between parietal and frontal cortex. Cerebral Cortex, 11(6),513–527.

Matelli, M., Govoni, P., Galletti, C., Kutz, D. F., & Luppino, G. (1998). Superior area 6afferents from the superior parietal lobule in the macaque monkey. Journal ofComparative Neurology, 402(3), 327–352.

Nudo, R. J., Milliken, G. W., Jenkins, W. M., & Merzenich, M. M. (1996). Use-dependent alterations of movement representations in primary motor cortex ofadult squirrel monkeys. Journal of Neuroscience, 16(2), 785–807.

Oostwoud Wijdenes, L., Brenner, E., & Smeets, J. B. (2011). Fast and fine-tunedcorrections when the target of a hand movement is displaced. ExperimentalBrain Research, 214(3), 453–462.

Oostwoud Wijdenes, L., Brenner, E., & Smeets, J. B. (2013). Analysis of methods todetermine the latency of online movement adjustments. Behavior ResearchMethods.

Pisella, L., Grea, H., Tilikete, C., Vighetto, A., Desmurget, M., Rode, G., et al. (2000). An‘automatic pilot’ for the hand in human posterior parietal cortex: Towardreinterpreting optic ataxia. Nature Neuroscience, 3(7), 729–736.

Prablanc, C., Desmurget, M., & Grea, H. (2003). Neural control of on-line guidance ofhand reaching movements. Progress in Brain Research, 142, 155–170.

Prablanc, C., & Martin, O. (1992). Automatic control during hand reaching atundetected two-dimensional target displacements. Journal of Neurophysiology,67(2), 455–469.

Proteau, L., Roujoula, A., & Messier, J. (2009). Evidence for continuous processing ofvisual information in a manual video-aiming task. Journal of Motor Behavior,41(3), 219–231.

Pruszynski, J. A., Kurtzer, I., Nashed, J. Y., Omrani, M., Brouwer, B., & Scott, S. H.(2011). Primary motor cortex underlies multi-joint integration for fast feedbackcontrol. Nature, 478(7369), 387–390.

1145


Reichenbach, A., Bresciani, J. P., Peer, A., Bulthoff, H. H., & Thielscher, A. (2011).Contributions of the PPC to online control of visually guided reachingmovements assessed with fMRI-guided TMS. Cerebral Cortex, 21(7), 1602–1612.

Reichenbach, A., Thielscher, A., Peer, A., Bulthoff, H. H., & Bresciani, J. P. (2009).Seeing the hand while reaching speeds up on-line responses to a sudden changein target position. Journal of Physiology, 587(Pt 19), 4605–4616.

Rice, N. J., Tunik, E., Cross, E. S., & Grafton, S. T. (2007). On-line grasp control ismediated by the contralateral hemisphere. Brain Research, 1175, 76–84.

Rice, N. J., Tunik, E., & Grafton, S. T. (2006). The anterior intraparietal sulcusmediates grasp execution, independent of requirement to update: New insightsfrom transcranial magnetic stimulation. Journal of Neuroscience, 26(31),8176–8182.

Robertson, D. G. E., & Dowling, J. J. (2003). Design and responses of Butterworth andcritically damped digital filters. Journal of Electromyography and Kinesiology,13(6), 569–573.

Sarlegna, F. R. (2006). Impairment of online control of reaching movements withaging: A double-step study. Neuroscience Letters, 403(3), 309–314.

Schluter, N. D., Rushworth, M. F., Passingham, R. E., & Mills, K. R. (1998). Temporaryinterference in human lateral premotor cortex suggests dominance for theselection of movements. A study using transcranial magnetic stimulation. Brain,121(Pt 5), 785–799.

Shadmehr, R., & Krakauer, J. (2008). A computational neuroanatomy for motorcontrol. Experimental Brain Research, 185(3), 359–381.

Soechting, J. F., & Lacquaniti, F. (1983). Modification of trajectory of a pointingmovement in response to a change in target location. Journal of Neurophysiology,49(2), 548–564.

Todorov, E., & Jordan, M. I. (2002). Optimal feedback control as a theory of motorcoordination. Nature Neuroscience, 5(11), 1226–1235.

Tunik, E., Frey, S. H., & Grafton, S. T. (2005). Virtual lesions of the anteriorintraparietal area disrupt goal-dependent on-line adjustments of grasp. NatureNeuroscience, 8(4), 505–511.

Vaziri, S., Diedrichsen, J., & Shadmehr, R. (2006). Why does the brain predict sensoryconsequences of oculomotor commands? Optimal integration of the predictedand the actual sensory feedback. Journal of Neuroscience, 26(16), 4188–4197.

Veerman, M., Brenner, E., & Smeets, J. (2008). The latency for correcting a movementdepends on the visual attribute that defines the target. Experimental BrainResearch.

Viviani, P., & Flash, T. (1995). Minimum-jerk, two-thirds power law, and isochrony:Converging approaches to movement planning. Journal of ExperimentalPsychology: Human Perception and Performance, 21(1), 32–53.

Wolpert, D. M., Ghahramani, Z., & Jordan, M. I. (1995). An internal model forsensorimotor integration. Science, 269(5232), 1880–1882.



visually-guided correction of hand reaching movements: the neurophysiological bases in the cerebral...

Documents