Unexpected changes in the location of a target for an upcoming action require both attentional reorienting and motor planning update. In both macaque and human brain, the medial posterior parietal cortex is involved in both phenomena but its causal role is still unclear. Here we used on-line rTMS over the putative human V6A (pV6A), a reach-related region in the dorsal part of the anterior bank of the parieto-occipital sulcus, during an attention and a reaching task requiring covert shifts of attention and planning of reaching movements toward cued targets in space. We found that rTMS increased RTs to invalidly cued but not to validly cued targets during both the attention and reaching task. Furthermore, we found that rTMS induced a deviation of reaching endpoints toward visual fixation and that this deviation was larger for invalidly cued targets. The results suggest that reorienting signals are used by human pV6A area to rapidly update the current motor plan or the ongoing action when a behaviorally relevant object unexpectedly occurs in an unattended location. The current findings suggest a direct involvement of the action-related dorso-medial visual stream in attentional reorienting and a more specific role of pV6A area in the dynamic, on-line control of reaching actions.
Effective goal-directed behavior depends on the ability to flexibly adapt a motor plan in response to unexpected changes of target location. Such motor reorganization requires that attention is rapidly shifted to the new spatial location even without an overt eye movement, a process typically defined as reorienting response (Corbetta, Patel, & Shulman, 2008). This adaptive response is typically associated with a right-lateralized ventral frontoparietal network (Corbetta et al., 2008; Corbetta & Shulman, 2002). Recent human neuroimaging studies, however, have shown that attentional reorienting also induces a transient increase of neural activity in regions of the medial superior parietal lobule extending from the precuneus to more posterior regions around the dorsal aspect of the parieto-occipital sulcus (POS; Tosoni, Shulman, Pope, McAvoy, & Corbetta, 2013; Shulman et al., 2009; Vossel, Weidner, Thiel, & Fink, 2009; Kelley, Serences, Giesbrecht, & Yantis, 2008; Molenberghs, Mesulam, Peeters, & Vandenberghe, 2007; Yantis et al., 2002). This portion of cortex is also specialized for visuo-motor coordination during arm-reaching movements (Galati et al., 2011; Striemer, Chouinard, & Goodale, 2011; Cavina-Pratesi et al., 2010; Vesia, Prime, Yan, Sergio, & Crawford, 2010; Busan, Monti, Semenic, Pizzolato, & Battaglini, 2009; Prado et al., 2005; Astafiev et al., 2003). In particular, the anterior region in the so-called superior parietal occipital cortex (SPOC) is robustly involved in both proximal (arm direction) and distal (hand orientation) aspects of reach-to-grasp movements (Monaco et al., 2011; Cavina-Pratesi et al., 2010).
The pattern of deficits exhibited by neuropsychological patients suffering from optic ataxia (OA; Karnath & Perenin, 2005; Perenin & Vighetto, 1988) offers further insights on the particular combination of attentional and motor functions of the dorso-medial parietal cortex. OA patients typically fail to make fast corrections of reaching movements when the target is unexpectedly displaced, suggesting a role of the medial parietal cortex in dynamic aspects of visual control of action (Glover, 2003; Gréa et al., 2002; Pisella et al., 2000). More importantly, the deficit in these patients is not confined to movements execution but also appears to affect the ability to detect and respond to targets located in the portion of the visual field (typically contralesional) in which the visuomotor deficit is most evident (McIntosh, Mulroue, Blangero, Pisella, & Rossetti, 2011; Striemer et al., 2007, 2009). Striemer et al. (2007, 2009) suggested that the deficits in attention and visuomotor control are independent (i.e., the two deficits result from damage to distinct mechanisms), whereas McIntosh et al. (2011) have recently suggested a single mechanism, showing that the visuomotor deficits observed in these patients and in particular their failure to use extrafoveal visual information to drive immediate actions could depend from an impairment in the ability to shift attention between visual locations.
Evidence from monkey neurophysiology has shown that V6A neurons in the dorsal POS, which are particularly sensitive to arm movements directed to nonfoveated objects (Marzocchi, Breveglieri, Galletti, & Fattori, 2008) and are modulated by gaze position (Galletti, Battaglini, & Fattori, 1995), also respond to covert shifts of attention (Galletti et al., 2010). In particular, Galletti and colleagues (2010) have shown that covert attentional modulations in V6A are consistent with the distribution of preferred reach direction in this area, suggesting that spatially directed attentional signals could be linked to arm motor programming.
To summarize, neuroimaging, neuropsychological, and neurophysiological evidence all converge to suggest a role of the cortex in the anterior dorsal part of the POS both in visuomotor transformations for goal-directed reaching movements and in attentional functions necessary to select salient or relevant information in the environment. It is unclear, however, whether action- and attention-related signals act independently (Striemer et al., 2007, 2009) or interact with each other in this cortical region (McIntosh et al., 2011; Galletti et al., 2010) and whether they are causally associated with its functioning.
To investigate these issues, we used a goal-standard technique of neural interference, such as TMS, which allows to draw causal inferences about the role of a brain region in a particular cognitive or sensorimotor function by inducing a “virtual lesion” in a restricted portion of the cerebral cortex (Bolognini & Ro, 2010; Paus, 2005; Pascual-Leone, Walsh, & Rothwell, 2000; but also see Casali, Casarotto, Rosanova, Mariotti, & Massimini, 2010, for a discussion of TMS effects on regions that are anatomically/functionally connected to the stimulated cortical sites). Individual MRI-guided TMS was carried out over a region in the dorsalmost part of the anterior bank of the POS (i.e., anterior SPOC) that likely corresponded to the human homologue of monkey area V6A (Pitzalis et al., 2012; Cavina-Pratesi et al., 2010) and that from now on we will refer to as the human putative area V6A (pV6A). We used a cueing paradigm in which participants performed both an attention and a reaching task. During the tasks, participants were asked to detect the appearance of a brief visual target presented in the peripheral visual field and to respond as quickly as possible with a right-hand button release only (attention task) or with a right-hand reaching movement to the target location (reaching task) while maintaining central fixation.
By manipulating the validity of the cues in both tasks (the cue correctly predicted target location with 75% probability), we tested the involvement of the pV6A in attentional reorienting during both the attention and the reaching task. To provide a comprehensive account of the TMS effects on the reorienting of attention toward unattended targets both when they have simply to be detected (attention task) and when they have to be processed as a goal of a reaching movement, we measured RTs to target detection in both tasks and the endpoints of reaching movements in the reaching task. We hypothesized that pV6A is causally involved in attentional reorienting, thus predicting a marked rTMS-induced increase of RTs for invalid trials in both tasks. Given that invalid targets, compared with valid ones, also require an update of the current motor plan, we additionally predicted a selective effect of validity on the pattern of rTMS-induced reaching errors.
Eight right-handed participants (four men, mean age = 26.1 years) participated in the experiment. All participants had normal or corrected visual acuity and reported no history of neuropsychiatric illness or epilepsy as well as any contraindication to TMS (Rossi, Hallett, Rossini, Pascual-Leone, & Group, 2009; Wassermann, 1998). All gave written informed consent in accordance with the guidelines of the local ethics committee and the ethical standards of the Declaration of Helsinki.
Stimuli and Apparatus
Participants were seated on a height-adjustable chair in complete darkness, with the head mechanically stabilized by a chin rest and a head holder mounted onto a wooden table positioned directly in front of them. A Plexiglas screen (120 × 50 cm) covered with a matte black sheet was located on the table within a reaching distance (35 cm). The height of the chair and the chin rest were adjusted so that the participant's cyclopean eye (located midway between the two eyes) was vertically and horizontally aligned with the central fixation light-emitting diode (LED). The stimuli array consisted of nine LEDs aligned to the horizontal plane: one green LED, located at 0°, served as fixation point, whereas the remaining eight yellow LEDs were located at four different eccentricities (5°, 10°, 20°, and 30°) on the left and the right of the central fixation LED and were used as cue and target stimuli. On each trial, a target was preceded by an informative peripheral cue of 10-msec duration, which correctly predicted the target location with a probability of 75% (valid trials). LEDs were installed behind the Plexiglas screen, were visible only when illuminated, and gave no tactile feedback when touched.
Eye position was monitored during both the attention and reaching task with an ISCAN ETL-400 remote infrared eye tracker (sampling rate: 120 Hz). Moreover, during the reaching task, the accuracy of reaching movements were recorded in all trials with an electromagnetic tracking device (Fastrak Polhemus digitizer, Polhemus; Colchester, VT). This electromagnetic tracking system provides measures of the position of small sensors attached to the tip of the right index fingers, with a sampling rate of 120 Hz and a spatial accuracy of 0.8 mm. Data were digitized and recorded on a PC for off-line analysis.
Individuation of Anatomical rTMS Sites
Cortical sites of rTMS stimulation were localized individually. To identify left and right pV6A stimulation sites within anterior SPOC and to monitor the TMS coil position at the end of each experimental block, we used a frameless stereotaxic neuronavigation system (Softaxic, EMS; Bologna, Italy). Before the experiment, a T1-weighted MR scan was obtained from each participant using a Siemens 3T scanner (1 × 1 × 1 mm, sagittal acquisition). Stimulation sites were then identified on the scalp by coregistering reference scalp locations to individual MR images using a neuronavigation system (Fastrak Polhemus) running a SofTaxic software. The pV6A sites on each hemisphere was localized according to individually determined anatomical landmarks as the region that is medial to the posterior end of the intraparietal sulcus (IPS) and anterior to the dorsal end of the POS (Figure 1A). This is the region where the human area V6A should be located. In fact, in the monkey, area V6A is located just anterior to V6 (Galletti, Fattori, Gamberini, & Kutz, 1999; Galletti, Fattori, Kutz, & Gamberini, 1999). Because in humans area V6 is located in the dorsalmost part of POS (Pitzalis et al., 2006), the human homologue of monkey V6A should be located just anterior to the dorsalmost part of POS, as also suggested by recent works (Pitzalis et al., 2012; Cavina-Pratesi et al., 2010). The target sites were marked on a tightly fitting Lycra cap worn by participants, and the coil was maintained in that position by an articulated metallic arm for the whole duration of the experimental block. The localization procedure was performed at the beginning of the experimental session and was controlled at the end of each experimental block.
Procedure for rTMS Stimulation
TMS was delivered via a 70-mm figure-of-eight induction coil, connected to a MagStim Rapid system (MagStim Company; Whitland, U.K.) and applied tangentially to the target scalp site, with the handle pointing posteriorly. The intensity of TMS was set at 60% of the stimulator output in accordance with previous studies on reaching- and attention-related activity in parietal cortex (Vesia et al., 2010; Dambeck et al., 2006). The TMS train consisted of three pulses (10 Hz) delivered at 0, 100, and 200 msec following the offset of the cue stimulus. The on-line rTMS train frequency, intensity, and duration were well within safe limits (Rossi et al., 2009; Wassermann, 1998).
The specificity of the behavioral effects of rTMS stimulation over left and right pV6A were controlled by including two control stimulation conditions in which rTMS was delivered, in equal number, over the same two sites but with the coil held perpendicular to the scalp (Sham) or over the Vertex (Cz according to the 10–20 EEG coordinate system).
Each experimental block started with central fixation (green LED; cross in Figure 1B). In each trial, a cue (yellow LED; filled gray circle in Figure 1B) was flashed for 10 msec in one of the eight possible locations along an horizontal line at eye level (i.e., four locations in each hemifield), at 5°, 10°, 20°, 30° of eccentricity, respectively. After an ISI of 240 msec, a target (yellow LED; filled black circle in Figure 1B) was presented either at the very same location (valid trials, 75%) or at a corresponding location in the opposite hemifield (invalid trials, 25%). Behavioral indices were measured during an attentional and a reaching task in which participants were instructed to indicate target detection by releasing a response button with the right index finger. In both tasks, we measured the RTs as the time between the onset of the target stimulus and the release of the response button. In the reaching task, participants were also required to perform a ballistic reaching movement to touch the location of the target on the screen.
Participants were informed about the task to perform (i.e., attentional or reaching task) at the beginning of each block. The blocks included 64 experimental trials (48 valid and 16 invalid) and six catch trials, in which the ISI was extended to 1000 msec to prevent early responses to the cue stimulus. Catch trials were not included in the subsequent analyses. To minimize TMS exposure and fatigue and to exclude learning effects, each participant completed the entire experiment in four sessions (performed in different days). Each session comprised six blocks, with task order and order of stimulation sites (both blocked) counterbalanced across sessions and individuals. Each participant completed 24 blocks (six blocks for each session) for a total of 1536 trials.
Gaze position (right eye) on the Plexiglas screen was recorded in each participant using an eye-tracking system, which was recalibrated before each block by means of a standard calibration procedure. Trials with eye blinks or with gaze deviation greater than 3° from central fixation (364 trials, corresponding to ≈3% of total trials) were discarded from the analyses of rTMS effects on behavioral performance. The dependent measures for both tasks were RTs, that is, the button release times. Trials with RTs shorter than 100 msec or longer than 1200 msec (626 trials, corresponding to 5% of total trials) were discarded from the analyses, because they were considered anticipatory or abnormally slow responses, respectively.
RTs were analyzed using a repeated-measure ANOVA with Task (attention, reaching), Stimulation Site (right pV6A, left pV6A, Sham/Vertex), Target Validity (valid, invalid), Visual Hemifield (left HF, right HF), and Target Eccentricity (±5°, 10°, 20°, 30°) as factors. Note that because no differences were found between the two control stimulation conditions (see Procedure for rTMS Stimulation section), they were collapsed in a single control condition (Sham/Vertex). When the sphericity assumption was violated, Greenhouse–Geisser corrected p values (indicated as pGG) were applied. The Newman–Keuls test was used for post hoc analyses. The threshold for statistical significance was set at p < .05 for all comparisons.
The accuracy of the reaching movements was evaluated by mapping the reaching endpoints on the horizontal (x) and vertical (y) axes of the screen. Endpoints were defined as the points at which the z position of the finger crossed the value that demarcates the location of the Plexiglas screen. We then estimated horizontal endpoint errors, measured in degrees of visual angle, as the signed difference between finger endpoint and target position as computed in the calibration procedure. In this calibration procedure, which was conducted at the end of the experiment, participants were requested to fixate and reach each LED targets with the full vision of their hand and without temporal constraints. We assessed the position of each target as the mean endpoint position averaged across five of these calibration reaching movements. This was done to take into account the possible small differences in the location of the electromagnetic sensor between participants. Reaching endpoints were analyzed by a repeated-measure ANOVA with the same factorial design described above, but without the Task factor.
We first describe the main effects and interactions emerged from the RT analysis that did not involve the Stimulation Site factor (Figure 2). As shown in Figure 2A, there was a significant effect of Validity, F(1, 7) = 20.54; p = .003, with longer RTs for invalid than valid trials (424 vs. 371 msec, respectively) and a significant effect of Target Eccentricity, F(3, 21) = 25.3; pGG < .0001, with progressively longer RTs as a function of target eccentricity (364, 386, 409, and 430 msec for targets located at 5°, 10°, 20°, and 30°, respectively; all post hoc comparisons were significant). Validity was also modulated by Target Eccentricity, F(3, 21) = 4.02; pGG = .036, with post hoc tests indicating a greater validity effect for targets located at 30° (66 msec) compared with other eccentricities (46, 52, and 50 msec for 5°, 10°, and 20°, respectively; all ps < .001). These results, obtained during control baseline stimulation, confirm that our paradigm was effective in generating a bias of spatial attention to the cued location that progressively increased as a function of eccentricity. Moreover, as shown in Figure 2B, the results indicated a significant effect of Task, with slower RTs during the reaching compared with the attention task (428 msec vs. 367 msec; F(1, 7) = 11.73; p = .011).
Finally, there was a significant Task × Visual Hemifield interaction, F(1, 7) = 10.5; p = .014, with post hoc tests showing longer RTs in the reaching task for target located in the left (431 msec) relative to the right (424 msec) visual hemifield, whereas no hemifield differences were observed in the attention task (364 and 370 msec for left and right hemifields, respectively). This finding suggests that RTs can be inflated by the additional processes required by movement planning, especially during reaching to the left hemifield, that is, the field contralateral to the responding hand.
TMS Effects on Attentional Reorienting
The main finding of our study was that rTMS over pV6A selectively affected attentional reorienting in both the reaching and the attention task (Figure 3). Specifically, we observed a significant, selective increase of RTs to invalidly cued targets during rTMS over both left and right pV6A compared with Sham/Vertex stimulation (Validity by Stimulation Site interaction: F(2, 14) = 8.58; p = .004). Specifically, although TMS did not affect RTs on valid trials (left pV6A: 372 msec and right pV6A: 372 msec vs. Sham/Vertex: 368 msec; ps > .6), we observed a virtually identical increase of RTs to invalid targets in the two tasks during TMS stimulation of pV6A compared with Sham/Vertex (left pV6A: 433 msec and right pV6A: 444 msec vs. Sham/Vertex: 394 msec; ps < .001). This result indicates that rTMS over pV6A specifically impairs attentional reorienting during invalid trials but not attentional orienting during valid trials.
TMS Effects on the Accuracy of Reaching Movements
Consistent with previous findings, the ANOVA on reaching endpoint errors (reaching task) indicated a significant decrease of horizontal hypermetria following rTMS stimulation of pV6A compared with Sham/Vertex (Vesia et al., 2010). In other words, the stimulation caused a deviation of reach endpoints toward visual fixation, thus reducing the classic tendency to reach too far peripherally relative to the central fixation point, a pattern of overshoot errors that we found in the Sham/Vertex condition and that is typically observed in reaching experiments. As shown in Figure 4, however, we also observed that this typical rTMS-induced decrease of hypermetria was modulated by the location of the visual target (Visual Hemifield × Stimulation Site interaction: F(2, 14) = 10.01; p = .002). Specifically, we found that, although rTMS over left pV6A versus Sham/Vertex induced a significant decrease of hypermetria errors during reaching movements directed toward both hemifields (left HF: −0.01° endpoint errors, p = .017; right HF: 0.41° endpoint errors, p = .038), rTMS over right pV6A versus Sham/Vertex only induced a significant decrease of endpoint errors for reaching movements directed toward left contralateral targets (left contralateral targets: endpoint errors = −0.20°, p = .013; right ipsilateral targets: endpoint errors = 0.83°, p = .065).
The ANOVA on reaching endpoints also revealed a significant Stimulation Site × Target Validity × Visual Hemifield interaction, F(2, 14) = 8.29; p = .004 (Figure 5). Post hoc analysis indicated that rTMS over pV6A led to a greater decrease of baseline hypermetria during Invalid than Valid trials, deviating reach endpoints even more toward visual fixation. In other words, the rTMS-induced reduction of reaching errors was stronger in invalid than valid trials, with this validity-dependent effect depending on the visual hemifield of target presentation (i.e., a greater rightward deviation for left targets and a greater leftward deviation for right targets). This validity-dependent, rTMS-induced effect was significant in all conditions with the exception of those in which stimulation was applied to left pV6A and targets were presented in the left hemifield (endpoint errors: 0.12° and −0.14° in Invalid and Valid trials, respectively, p = .12).
In this study, we used on-line rTMS to test the relationship between attentional modulations and reaching movement execution in a region of the anterior SPOC that likely corresponds to the human homologue of monkey area V6A (Pitzalis et al., 2012; Cavina-Pratesi et al., 2010), an area known in the macaque to have both reaching and attentional responses (Galletti et al., 2010; Fattori, Kutz, Breveglieri, Marzocchi, & Galletti, 2005). We designed two tasks requiring participants to detect peripheral visual targets, which were either validly or invalidly cued, and to respond with a simple button release (attention task) or with a button release followed by a ballistic reaching movement toward the target location (reaching task). In both tasks, we evaluated the effect of rTMS stimulation (pV6A vs. control Sham/Vertex) on attentional components of orienting versus reorienting to target location. In the reaching task, we further evaluated the effect of TMS on reaching kinematic (endpoint errors).
The results of the RTs analysis on simple behavioral effects first indicated that our cueing paradigm was effective in generating a bias of spatial attention to the cued location in both tasks (i.e., the validity effect; Posner, Snyder, & Davidson, 1980). Importantly, the validity effect increased as a function of target eccentricity with progressively longer RTs for invalid targets presented at greater eccentricities. This is in line with a widely accepted assumption of dominant models of attention, which posits that the longer the path of attentional shifting, the greater the response delay for target detection (Hamilton, Stark, & Coslett, 2010; Henderson & Macquistan, 1993).
The main result of the study is the finding of a selective rTMS-induced increase of RTs for invalid trials during both tasks. Consistent with recent neurophysiological findings (Galletti et al., 2010), these results demonstrate that pV6A does not simply participate in the sensory-motor transformations needed to encode reach goals during goal-directed actions but also encodes critical signals for shifts of spatial attention. In particular, our findings suggest that pV6A plays a causal role in attentional reorienting, that is, when attention must be disengaged from one peripheral location and redirected to another peripheral location, but not in attentional orienting. These findings are in agreement with recent neuroimaging results showing that signals for shifting attention between peripheral locations, that is, reorienting, are specifically encoded in the medial aspect of the superior parietal cortex (Tosoni et al., 2013; Vossel et al., 2009; Kelley et al., 2008; Molenberghs et al., 2007; Serences & Yantis, 2006; Yantis et al., 2002; Wojciulik & Kanwisher, 1999), whereas spatially selective signals for maintaining attention at a location, that is, orienting, are encoded in more lateral parietal regions (Serences & Yantis, 2007; Corbetta & Shulman, 2002).
Importantly, our results also indicate that stimulation of pV6A during invalid condition is not associated with a modulation of RT performance that depends on target eccentricity (Stimulation Site × Target Validity × Target Eccentricity interaction: p > .5). Although this is a null result, it is an intriguing observation. One possible interpretation is that pV6A is specifically implicated in the disengagement phase of reorienting rather than in the subsequent shift to the new location (Posner, Walker, Friedrich, & Rafal, 1984). In fact, a region that specifically supports signals for shifting attention between peripheral locations would be expected to show a linear increase of invalid RTs as target eccentricity increases, and this was not the case.
In the reaching task, the analysis of movement accuracy revealed that stimulation of pV6A significantly reduced the reaching “overshoot” errors by deviating reach endpoints toward visual fixation. This resembles the “magnetic misreaching” found in OA patients (Carey, Coleman, & Della Sala, 1997). In particular, we observed that reaching hypermetria, the classic tendency to reach too far peripherally relative to the central fixation point, which is typically observed in behavioral experiments involving spatially guided reaching movements (Ambrosini et al., 2012; Medendorp & Crawford, 2002; Bock, 1986), was reduced following pV6A compared with Sham/Vertex stimulation. Such rTMS-induced effect on reaching kinematics is consistent with previous findings (Vesia et al., 2010), indicating that our stimulation sites effectively involved a reach-related cortical region.
The present rTMS results on reaching performance are also consistent with the visuomotor deficits observed in OA patients with unilateral posterior parietal lesions (Perenin & Vighetto, 1988). We indeed found that inactivation of right pV6A (ipsilateral to the responding hand) affected reaching accuracy only in the contralateral visual hemifield, resembling the so-called “field effect” observed in OA patients (Perenin & Vighetto, 1988) and suggesting that reach accuracy is influenced by visual hemifield (Ciavarro & Ambrosini, 2011). Differently from right pV6A inactivation, the left pV6A inactivation (contralateral to the responding hand) impaired reaching movements directed to targets in both hemifields, an effect that is also reminiscent of the so-called “hand effect” typically observed in OA patients (Blangero et al., 2010). However, as we required participants to perform reach movements with the right hand only, further investigations are needed to confirm these conclusions.
Moreover, consistent with the selective effect of pV6A stimulation on RTs during invalid trials, we found that rTMS over pV6A, compared with baseline, induced a greater reduction of baseline hypermetria when reaching movements were directed toward invalidly cued targets. Therefore, pV6A stimulation during invalid trials was associated with both an overall increase of RTs at target detection (in both attention and reaching tasks) and with an increment of reaching endpoint errors. This suggests a tight functional link between reaching and attentional processes when attention is reoriented from one location to another, as if shifts of attention are necessary for the corresponding update of reaching target. This result is consistent with recent neurophysiological data (Galletti et al., 2010) showing modulations for covert spatial attention in monkey area V6A, although a direct comparison between the studies is complicated by the absence in neurophysiological recordings of conditions in which monkeys are trained to reorient attention and execute reaching movements toward unattended targets. Galletti and colleagues (2010) showed that covert attention modulations in area V6A are consistent with the distribution of preferred gaze and reach direction observed in that area, rather than with the distribution of visual receptive fields (that in V6A are mainly located in the contralateral visual field), suggesting that attentional and reach activity are closely related in that cortical area.
Importantly, our findings can help disentangling different hypotheses about the link between attentional and visuomotor deficits in OA patients (McIntosh et al., 2011; Striemer et al., 2009). For example, McIntosh et al. (2011) have suggested that the two deficits could be linked because peripheral target jumps slowed perceptual discrimination and mirrored the reaching deficit. Although the experimental tasks in this neuropsychological study were mainly designed to test specific deficits associated with reaching on-line correction in OA patients, the findings are in line with our results. Our results, however, are in contrast with those of Striemer and colleagues (2009), who did not find a common pattern of errors between attention and reaching tasks in OA patients compared with the control group, thus proposing that attentional and visuomotor deficits arise from independent mechanisms. However, it is worth noting that in this study the authors have not used a cued paradigm and compared very different behavioral measures (RTs and reaching accuracy), which does not represent an optimal basis to contrast the performance between attention processes and planning of arm-reaching movements.
It should be noted here that the issue of attentional and reaching functions in parietal cortex has been already addressed in two recent TMS studies. In particular, in the study by Vesia and colleagues (2010), rTMS was used to determine effector specificity (spatially directed reaching and saccadic eye movements) in the posterior parietal cortex. One of the main findings was that stimulation of SPOC did not affect saccadic eye movements but deviated reach endpoints toward visual fixation. This result is entirely consistent with our findings of a significant decrease of horizontal hypermetria following stimulation of pV6A, a region that is included in the SPOC. Compared with our work, however, the study by Vesia and colleagues (2010) did not address the question of attentional modulations in the reach-related SPOC region. Attentional effects associated with target spatial validity were instead investigated in the series of studies by Capotosto and colleagues (Capotosto, Babiloni, Romani, & Corbetta, 2009, 2012; Capotosto, Corbetta, Romani, & Babiloni, 2012), who employed a Posner-like task to examine TMS interference on EEG rhythms and behavioral performance during spatial orienting and reorienting. As in our work, Capotosto and colleagues (Capotosto, Babiloni, et al., 2012; Capotosto, Corbetta, et al., 2012; Capotosto et al., 2009) observed that TMS more strongly impaired performance during invalid than valid trials (note however that also a significant TMS effect on valid trials was observed in Capotosto, Corbetta, et al., 2012). One important difference, however, concerns the location of stimulation sites. Specifically, whereas Capotosto and colleagues applied TMS to parietal regions in the posterior IPS, we targeted the putative human V6A region, which is located in the anterior bank of the dorsalmost POS (i.e., anterior SPOC) and thus more medial and posterior than the IPS by Capotosto (note that the estimated distance between the two cortical sites is ≈4 cm, that is, beyond the spatial resolution of the TMS; Wagner, Valero-Cabre, & Pascual-Leone, 2007). Other notable differences include the use in our study, but not in that of Capotosto et al. (Capotosto, Babiloni, et al., 2012; Capotosto, Corbetta, et al., 2012; Capotosto et al., 2009), of peripheral stimuli at different visual eccentricity, which allowed to test specific TMS effects on visual representations, and the combined evaluation of TMS effects on attentional and action-related (i.e., reaching) functions in pV6A.
To sum up, our findings represent both a confirmation and an extension of available data about attentional and reaching functions in the medial posterior parietal cortex. They are in line with current proposals of a functional segregation between medial and lateral regions of the superior parietal lobule for attention (Capotosto et al., 2013) and reaching (Vesia & Crawford, 2012) processes. In particular, whereas lateral areas, including the IPS, would encode spatially selective signals for attending to a location (Corbetta & Shulman, 2002) and motor details for the reach vector (Vesia et al., 2010), our findings suggest that medial areas, including pV6A, are specialized for encoding signals for shifting attention between peripheral locations (e.g., Yantis et al., 2002) as well as to peripheral reach goals (e.g., Vesia et al., 2010).
On the basis of the findings reported in this article, we propose that reorienting signals are used by the human pV6A to rapidly update the current motor plan or the ongoing action when a behaviorally relevant object unexpectedly appears at an unattended location, requiring a rapid and adaptive motor response such as reaching, grasping or pushing it away. On this basis, we suggest a direct involvement of the action-related dorso-medial visual stream in attentional reorienting and a more specific role of pV6A area in the dynamic, on-line control of reaching actions.
This work was supported by grants from MIUR to G. C. (prot. 006055034_002) and C. G. (prot. 2006055034_001), Fondazione del Monte di Bologna e Ravenna, Italy, to C. G., and Fondazione Neurone Onlus Roma, Italy, to M. C.
Reprint requests should be sent to Prof. Claudio Galletti, Department of Pharmacy and Biotechnology, University of Bologna, Piazza di Porta San Donato, 2, 40126, Bologna, Italy, or via e-mail: firstname.lastname@example.org.