Spatial attention can be defined as the selection of a location for privileged stimulus processing. Most oculomotor structures, such as the superior colliculus or the FEFs, play an additional role in visuospatial attention. Indeed, electrical stimulation of these structures can cause changes in visual sensitivity that are location specific. We have proposed that the recently discovered ocular proprioceptive area in the human postcentral gyrus (S1EYE) may have a similar function. This suggestion was based on the observation that a reduction of excitability in this area with TMS causes not only a shift in perceived eye position but also lateralized changes in visual sensitivity. Here we investigated whether these shifts in perceived gaze position and visual sensitivity are spatially congruent. After continuous theta burst stimulation over S1EYE, participants underestimated own eye rotation, so that saccades from a lateral eye rotation undershoot a central sound (Experiment 1). They discriminated letters faster if they were presented nearer the orbit midline (Experiment 2) and spent less time looking at locations nearer the orbit midline when searching for a nonexistent target in a letter array (Experiment 3). This suggests that visual sensitivity increased nearer the orbit midline, in the same direction as the shift in perceived eye position. This spatial congruence argues for a functional coupling between the cortical eye position signal in the somatosensory cortex and visuospatial attention.
Gaze and attention are tightly coupled. This observation during everyday behavior is substantiated by the shared neuroanatomical substrates and the functional interactions between the oculomotor and visuospatial attention systems (Smith & Schenk, 2012; Rizzolatti, Riggio, Dascola, & Umiltá, 1987). Stimulation of the oculomotor structures, such as the FEFs or the superior colliculi, can not only evoke saccades (Bruce, Goldberg, Bushnell, & Stanton, 1985; Robinson, 1972) but can also create areas of increased visual sensitivity (Ruff et al., 2006; Cavanaugh & Wurtz, 2004; Moore & Fallah, 2004). These studies have established a link between the oculomotor structures and visuospatial attention. A recently discovered oculomotor structure is the eye proprioceptive area in the human somatosensory cortex (S1EYE; Balslev, Himmelbach, Karnath, Borchers, & Odoj, 2012; Balslev & Miall, 2008), presumably a homologue of the eye proprioceptive area in the monkey (Wang, Zhang, Cohen, & Goldberg, 2007). We have previously suggested that S1EYE, in addition to its role in coding the rotation of the eye in the orbit, also has a function in visuospatial attention (Balslev, Siebner, Paulson, & Kassuba, 2012; Balslev, Gowen, & Miall, 2011). This suggestion was based on the following observation. Interfering with the function of S1EYE by using inhibitory repetitive TMS caused lateralized changes in visual sensitivity that depended on the direction of eye rotation (Balslev et al., 2011). For instance, when the right eye was rotated leftward, 1 Hz rTMS over S1EYE decreased the accuracy for detecting left compared with right visual hemifield targets. When the right eye was rotated rightward, we found the opposite results; now the right hemifield targets were detected less accurately than those on the left. These changes in visual sensitivity were mirrored by changes in the activity of the extrastriate visual cortex after S1EYE-rTMS (Balslev, Siebner, et al., 2012).
To further characterize the function of the somatosensory cortex in visuospatial attention, here we tested the hypothesis that the lateral shift in visual sensitivity induced by changing the excitability of S1EYE is spatially congruent with the rTMS-induced shift in perceived gaze direction.
To this end, we tested both perceived direction of gaze and visual sensitivity for two directions of eye rotation, leftward and rightward, before and after applying continuous theta burst stimulation (cTBS) over S1EYE in healthy participants. cTBS decreases the excitability of the area underneath the coil when applied to the motor (Huang, Edwards, Rounis, Bhatia, & Rothwell, 2005) or the somatosensory (Ishikawa et al., 2007) cortices. We measured the effect of S1EYE-cTBS on perceived eye position (indexed by the amplitude of the saccades from a lateral fixation point to a central auditory target, Experiment 1) as well as on visual sensitivity (indexed by the RT for visual discrimination, Experiment 2).
In a third experiment, we asked whether S1EYE-cTBS also causes a lateral bias in visual exploration (Experiment 3). For instance in the posterior parietal cortex, a decrease of excitability of the right posterior parietal cortex with inhibitory rTMS not only increases visual sensitivity in the right hemifield (Hilgetag & Pascual-Leone, 2001) but also increases the time the participants spend looking in the right hemispace when viewing visual scenes (Nyffeler et al., 2008).
All participants were healthy adults, right-handed by self-report, with normal, or corrected-to-normal vision. They gave written, informed consent to participate. The study was approved by the Ethics Committee of the Medical Department of the University of Tuebingen. For Experiment 1, 14 participants were tested: age median = 25 years, range = 20–28 years, 10 women. In Experiment 2, we tested 13 participants: age median = 28 years, range = 23–33 years, 7 women. For Experiment 3, we tested 12 participants: age median = 28.5 years, range = 23–33, 5 women. Two participants took part in all three experiments, five participants took part in Experiments 1 and 2, and three participants participated in Experiments 2 and 3.
All tasks were performed in right, monocular vision while the left eye was covered with a patch. The head was fixed in a chin rest using cheek pads. The position of the right eye was recorded.
This experiment investigated the effect of S1EYE-cTBS on perceived eye position by measuring the saccade amplitude from a lateral, visual fixation point to a central auditory target. As a control, a second condition where the target was presented in the visual rather than auditory modality was used. The main oculomotor structures such as the FEF (Russo & Bruce, 1994) or the superior colliculi (Jay & Sparks, 1984) code the direction of the saccade for both auditory and visual targets in an eye-centered reference frame. Saccades to visual targets can be planned in an eye-centered reference frame without information about eye position at saccade onset. In contrast, auditory targets are initially coded relative to the ears and the head. Therefore, knowledge of eye position is critical for planning saccades to sounds.
The participants sat in a completely dark room. An LED array was placed at 46 cm in front of them, with its center aligned to the head and body midline. LEDs indicated fixation points at 10° to the left or right as well as the visual target at 0° (straight ahead). The auditory target was a 100-msec burst of white noise presented in a small speaker (diameter = 0.5 cm) attached to the LED array at 0° (Figure 1A, B). At the beginning of each trial the fixation LED was lit for 1100 msec. Then the target (LED or sound) was presented for 100 msec, then both target and fixation were turned off. The participant was instructed to move the eyes to the fixation LED and maintain fixation until both fixation and target were turned off, then to saccade to the location of the target. When ready for the next trial, the participant informed the experimenter, who started the next trial by pushing a button. Each run consisted of a block of 40 trials to the auditory target and 20 trials to the visual target. The order of visual and auditory blocks were identical within participants for both cTBS runs (pre- and post-cTBS) and stimulation sites (S1EYE and M1) and randomized across participants. Before the experiment, all participants underwent a practice run with eight trials per block. The training blocks were in the same order as the experiment blocks and were repeated until the participants performed at least 75% correctly timed saccades in each block.
Data were analyzed with repeated-measure ANOVAs. Post hoc pairwise comparisons of the performance before and after cTBS were conducted using the Tukey's test. This test takes into account the multiple comparisons.
This experiment aimed to test whether S1EYE-cTBS causes lateral shifts in visual sensitivity. To this end, we attempted to replicate the observations of Balslev and colleagues (Balslev et al., 2011) using a TMS method and a setup that was consistent with Experiment 1. Thus, we used cTBS instead of 1-Hz rTMS and a setup where the gaze was maintained to the left or to the right, symmetrically relative to the orbit midline.
A computer screen was positioned with its center at ±14° visual angle from the orbit midline of the right eye (Figure 2A, B). At the start of the trial, a fixation cross (1° × 1°) was presented at the center of the screen (14° left or right from orbit center) for a randomized time interval between 1600 and 2400 msec. Then a target letter (“A” or “H,” 1° × 1°) appeared for 120 msec at 10° visual angle to the left or to the right of the screen center. The participant was instructed to fixate the cross and respond as fast as possible to indicate the letter. To respond, he or she pressed with the left hand one of the two buttons of the response box (“A” or “H”). The participant had 2000 msec to respond before the start of a new trial. We measured the RT from target onset until button press as an index for visual sensitivity. RT for visual stimuli decreases with increasing stimulus luminance (Jaskowski, Rybarcyzk, Jaroszyk, & Lemanski, 1995) and is therefore an adequate measure for visual sensitivity. Each participant completed 160 trials: the first 80 with the screen on one side and the last 80 trials with the screen on the other side. The order of the two gaze conditions (leftward, rightward) was counterbalanced across participants. Within each gaze condition, there were 20 trials for each letter and side of target presentation. The four different types of trials were intermixed in a pseudorandomized order. Before the experiment started, each participant underwent a practice run with 30 trials. To verify fixation, we recorded eye position.
Data were analyzed with repeated-measure ANOVAs. Post hoc pairwise comparisons of the performance before and after cTBS were conducted using Tukey's tests. These tests took into consideration all pairwise comparisons for each side of the stimulus display.
To investigate the pattern of exploratory eye movements during visual search, participants were instructed to look for the letter “A” within a letter array (height × width, 50° × 25°). The array was presented laterally at 14° visual angle to the left or to the right of the center of the right orbit so that, to fixate any letter within the array, the eye had to be rotated leftward or rightward, respectively (Figure 3A, B). Each array showed randomly distributed black capital letters from B to Z (1° visual angle, 113 letters in the left and 119 letters in the right array). Each letter was printed on a separate piece of white paper, attached to the array. Unbeknownst to the participants, the letter “A” was absent from the array. This paradigm for investigating the exploratory, spontaneous eye movements during visual search was adapted from Karnath and colleagues (Karnath, Niemeier, & Dichgans, 1998). Before the trial, the arrays were covered. At the beginning of the trial, the cover was removed and the participants instructed to search one lateral array. Eye movements were recorded for 20 sec, then the array was again covered. The mean gaze position during this search was calculated and compared across conditions.
Data were analyzed with repeated-measure ANOVAs. Post hoc pairwise comparisons of the performance before and after cTBS were conducted using Tukey's tests, which took into consideration all pairwise comparisons for each side of the letter array. To keep the participants motivated, in a practice trial before the experiment, a letter A was attached at the center of the left array. If the participant did not find the target within 20 sec, the experimenter identified its location by pointing to it. This was the case in one participant. Each participant completed two trials of 20-sec duration, one for each lateral array (leftward and rightward).
The position of the right eye was recorded with a head-mounted tracker (EYELINK II, SR Research Ltd., Ottawa, Canada) that sampled pupil location at 250 Hz. The tracker was calibrated after each cTBS run (pre- or post-cTBS) and for each lateral position of the screen (Experiment 2) or array (Experiment 3) using a horizontal grid (Experiment 1) or a 3 × 3 grid (Experiments 2 and 3).
Eye position time series were parsed into fixations, blinks, and saccades using the SR EyeLink detection algorithm, which was set to detect saccades with an amplitude of at least 0.5°, using an acceleration threshold of 9500/sec2 and a velocity threshold of 30/sec, and then analyzed off-line. To assess the effect of cTBS on calibration, we compared its accuracy and precision between the pre- and post-cTBS runs using paired t test. Calibration accuracy was indexed by the signed error between the position of the calibration point and the eye position returned by the calibration algorithm. Precision was measured by the standard deviation of this error.
For Experiment 1, trials with anticipatory saccades, performed before the target was extinguished, were discarded. For the remaining trials, we calculated the amplitude of the first saccade after the target was turned off. In addition, to investigate whether S1EYE-cTBS changed the initial eye position during fixation, we compared eye position averaged over 200 msec of fixation before saccade onset across conditions (pre- vs. post-cTBS) using paired t tests. To assess the effect of cTBS on the ability to plan and execute saccades, we compared saccade velocity and RT until saccade onset between the two cTBS runs (pre- vs. post-cTBS).
For Experiment 2, mean gaze position was calculated for a 200-msec interval before the target onset. Trials with a gaze deviation of more than 1° from the fixation cross were discarded. To investigate whether cTBS caused a lateral bias in the spontaneous eye movements, we calculated the amplitude and frequency of incidental saccades during required fixation and submitted them to the same repeated-measures ANOVA as the RT data.
For Experiment 3, blinks were excluded, and then mean eye position during visual search was calculated. To investigate whether changes in mean eye position were caused by lateralized changes in the fixation time or the amplitude or frequency of the saccades, these parameters were calculated and compared across conditions using repeated-measures ANOVAs and Tukey's tests.
A standard 70-mm-diameter figure-of-eight coil centered over the stimulation site was fixed in place by a coil holder. The participant's head was restrained by a chin rest. We followed an identical procedure for locating the eye proprioceptive area as in previous studies, conducted by Balslev and colleagues (Balslev et al., 2011; Balslev & Miall, 2008). S1EYE was mapped in each participant in relation to the “motor hotspot” of the left hemisphere, which is the scalp projection of the primary motor cortex for the hand (M1). The motor hotspot was defined as the point of maximum evoked motor response in the first dorsal interosseus muscle of the right hand. The S1EYE site of stimulation was located at 3 cm posterior to the motor hotspot, measured on a line oriented at 45° from the sagittal plane and perpendicular on the central sulcus. Post hoc neuronavigation has showed that this coil location targets the postcentral gyrus (Figure 2 in Balslev, Siebner, et al., 2012). During stimulation, the coil was positioned tangential to the scalp with the long axis of the figure-of-eight coil oriented at 45° to the parasagittal plane. The current flow of the initial rising phase of the biphasic pulse in the TMS coil induced a current flowing from posterior to anterior in the brain. On the basis of the decreased amplitude of the somatosensory-evoked potentials after cTBS over a region situated at 2 cm posterior to M1 (Ishikawa et al., 2007), we assumed that cTBS over S1EYE results in a decreased excitability of this area. The control site of stimulation M1 was located at the motor hotspot for the right hand. The choice of this control area was motivated by our previous TMS studies showing no effect of inhibitory TMS at this site on either visual localization (Balslev & Miall, 2008) or visuospatial attention (Balslev et al., 2011).
Each cTBS consisted of 600 biphasic stimuli produced by a Magstim Rapid2 stimulator. They were delivered with a frequency of three pulses at 50 Hz repeated at 200 msec (5 Hz) for 40 sec. The stimulation intensity was set at 80% of active motor threshold of the right first dorsal interosseus (Huang et al., 2005). For each experiment, participants underwent two sessions, with cTBS at either S1EYE or the control site (M1). The order of the sessions was randomized across participants and scheduled on separate days. During each session, the participant was tested before (pre-cTBS) and after (post-cTBS) on an identical task. Data collection was completed within 13 min after the cessation of the stimulation, a time interval for which the inhibitory aftereffect of cTBS in the somatosensory cortex has been demonstrated (Ishikawa et al., 2007). To prevent a potential recalibration of eye position by vision (Duke, Oruç, Qi, & Backus, 2006), participants were instructed to keep their eyes closed while receiving cTBS and until the start of the post-cTMS run.
This experiment investigated whether cTBS over the somatosensory cortex causes an underestimation of the rotation of the eye in the orbit. We predicted that S1EYE-cTBS would reduce the amplitude of the saccades from a lateral fixation point to a central auditory target. Because saccades to visual targets can be planned in retinotopic coordinates without information about eye position, no effect of S1EYE-cTBS on the amplitude of the visually guided saccades was expected.
A total of 16.18 ± 8.90% (mean ± standard deviation) trials were discarded because of anticipatory saccades, which started before the target was extinguished. As predicted, after S1EYE-cTBS, saccades from a lateral fixation to a central auditory were ∼1 degree shorter compared with the pre-cTBS and the M1-cTBS control, causing errors to the left for leftward fixation and to the right for rightward fixation (Figure 1C–F). Indeed for leftward eye rotation at the saccade onset, the amplitude of the saccades was 9.45° ± 3.15° before and 8.59° ± 3.01° after cTBS. For rightward eye rotation, the corresponding values were 9.48° ± 2.20° pre-cTBS and 8.32° ± 1.94° post-cTBS (pairwise comparisons using Tukey's test, both p < .05). The reduction in saccade amplitude post-cTBS versus pre-cTBS was specific for S1EYE and the auditory targets. None of the other comparisons (e.g., for M1-cTBS or for visual targets) were statistically significant (all ps > .05).
The repeated-measures ANOVA with factors (1) Target Modality (visual or auditory), (2) Stimulation Site (M1 or S1EYE), (3) cTBS Run (pre- or post-cTBS), and (4) Side of Saccades (left or right) showed a significant four-way interaction, F(1, 13) = 9.452, p = .009. A significant interaction could be found between Modality × Run × Side of Saccades, F(1, 13) = 13.09, p = .003. Further analyses (three-way ANOVAs with factors Modality × Run × Side of Saccades ran separately for S1EYE-cTBS and M1-cTBS data) showed a significant three-way interaction in the S1EYE-cTBS only (S1EYE-cTBS, F(1, 13) = 24.88, p < .001; M1-cTBS: F(1, 13) = 0.006, p = .94). None of the other main effects or interactions was significant.
We found no evidence that S1EYE-cTBS interferes with the participants' ability to fixate. The repeated-measures ANOVA for initial eye position data showed no significant three-way interaction between Stimulation Site × cTBS Run × Side of Saccades for either visual or auditory targets (both p > .53). Likewise, post hoc pairwise comparisons using the Tukey's test showed no significant difference of initial eye position pre-cTBS versus post-cTBS in either condition (M1-cTBS or S1EYE-cTBS for visual or auditory targets).
Furthermore, the effect of S1EYE-cTBS was specific to the amplitude of the saccades and did not extend to other measures of saccade kinematics. The analysis of saccade velocity or RT until saccade onset showed no significant effect of cTBS at either stimulation site and for either of the two modalities of the target (ANOVAs with factors Stimulation Site × cTBS Run × Side of Stimulation for saccade velocity and RT for both modalities: all ps > .261 for main effects and interactions).
The reduction in amplitude of the saccades to auditory targets after S1EYE-cTBS indicated an error in saccade planning caused by an underestimation of the rotation of the eye in the orbit. S1EYE-cTBS can thus dissociate the real direction of gaze from the perceived direction of gaze. Thus, it provides a method to investigate whether changes in perceived eye position are associated with changes in visual sensitivity.
Experiment 2 used the cTBS method to investigate for a lateral change in visual sensitivity. To this end, we measured RTs for discriminating briefly presented letters in the left and right visual hemifield in leftward or rightward gaze. If visual sensitivity shifts in the same direction as perceived eye position does, S1EYE-cTBS should speed up the discrimination of the letters located nearer the orbit center (in the left visual hemifield in rightward gaze and in the right visual hemifield for leftward gaze).
Trials with a break of fixation were discarded (8.53 ± 3.13%). In accord with the prediction, S1EYE-cTBS (but not M1-cTBS) changed the left–right gradient in the RT for letter discrimination to favor locations nearer the center of the orbit (Figure 2C, D). That is, in leftward gaze, the difference in RT for left minus right targets increased (Figure 2C, pre-S1EYE-cTBS: 1.22 ± 19.64 msec, post-S1EYE-cTBS: 16.33 ± 25.26 msec, post hoc pairwise comparisons, Tukey's test, p < .05). In rightward gaze, the difference in RT for left minus right targets decreased (Figure 2D, pre-S1EYE-cTBS: 7.53 ± 14.09 msec, post- S1EYE-cTBS: −3.07 ± 13.31 msec, Tukey's test, p < .05). Similar pairwise comparisons for the RTs after M1-cTBS showed no statistically significant (all ps > .05). The repeated-measures ANOVA with factors (1) Stimulation Site (S1EYE or M1), (2) cTBS Run (pre-cTBS or post-cTBS), (3) the Direction of Rotation of the Right Eye (leftward or rightward), and (4) Visual Hemifield in which the target letter appeared (left or right) showed a significant four-way interaction, F(1, 12) = 13.58, p = .003. A significant interaction was also found between Stimulation Site × the Direction of Rotation of the Right Eye × Target Side, F(1, 12) = 6.474, p = .023. This interaction was driven by post-cTBS data. Three-way ANOVAs (Stimulation Site × Direction of Rotation of the Right Eye × Target Side) was significant for post-cTBS-data, F(1, 12) = 16.996, p = .001, but not pre-cTBS data, F(1, 12) = 0.112, p = .74. No other main effects or interactions were found in any of the ANOVAs.
Discrimination accuracy over all participants and conditions approached ceiling (94.52 ± 5.07%) with no significant difference between conditions (repeated-measures ANOVA, four-way interaction, F(1, 12) = 0.605, p > .45). We found a nonsignificant trend (p = .094) for an interaction between stimulation site and eye rotation, which is irrelevant for the tested hypothesis. We failed to find any other statistical significant main effect or interaction (all ps > .15).
In contrast with the RT for letter discrimination, the analysis of incidental saccades showed no statistically significant spatial bias in the eye movements after S1EYE-cTBS. The four-way interaction was not significant for either the amplitude, F(1, 12) = 1.227, p = .29, or for the frequency, F(1, 12) = 0.334, p = .57, of these incidental saccades. No significant main effects or interactions were found.
The results of Experiment 2 confirmed the hypothesis that S1EYE-cTBS shifted visual sensitivity toward the center of the orbit, in the same direction as the shifts in perceived eye position (Experiment 1). We wondered whether this reflects a more general bias for the space nearer the center of the orbit, for instance, whether in addition to the increase in visual sensitivity, the exploratory eye movements too will show a preference toward this space.
To investigate whether S1EYE-cTBS causes a lateral shift in the exploratory eye movements, participants performed a visual search task in lateral letter arrays while their gaze position was recorded.
We found that after S1EYE-cTBS mean gaze during visual search was shifted ∼1° away from the orbit midline, that is to the left during visual search in the leftward letter array (pre-S1EYE-cTBS: −12.35 ± 1.4°, post-S1EYE-cTBS: −13.36 ± 2.21°, post hoc pairwise multiple comparison using Tukey's test, p < .05; Figure 3C) and to the right for the rightward array (pre-S1EYE-cTBS: 13.57 ± 1.28°, post-S1EYE-cTBS: 14.91 ± 2.77°, post hoc pairwise multiple comparison using Tukey's test p < .01; Figure 3D). This effect was specific to S1EYE-cTBS. After M1-cTBS, there was no statistically significant difference in mean gaze position during visual search in either array. The repeated-measures ANOVA with factors (1) Stimulation Site (S1EYE or M1), (2) cTBS Run (pre-cTBS or post-cTBS), and (3) the Side of the Array (left or right) showed a statistically significant three-way interaction, F(1, 11) = 17.66, p = .001. A significant main effect was found for eye rotation, F(1, 11) = 1507.51, p < .001. This main effect can be explained by the different direction of eye rotation for each array, giving negative values for rotation to the left and positive values for rotation to the right. None of the other main effects and no interactions were significant.
The shift in mean eye position after cTBS was caused by an increase in the fixation time in the “outer” half of the array (the half of the array located further from the orbit midline), rather than by an increase in the amplitude or the frequency of the saccades in that direction. Indeed, S1EYE-cTBS increased the left minus right gradient in the fixation time for the left array (pre-S1EYE-cTBS: −4.08 ± 2.46 sec, post-S1EYE-cTBS: −1.78 ± 3.14 sec, post hoc pairwise comparisons, Tukey's test, p < .05). In the right array, the left minus right gradient shifted toward the right half of the array (pre-S1EYE-cTBS: 1.68 ± 2.54 sec, post-S1EYE-cTBS: −0.68 ± 1.66 sec, Tukey's test, p < .05). Similar pairwise comparisons for fixation time after M1-cTBS showed no statistically significant difference (pre-M1-cTBS for left array: −3.01 ± 2.77 sec, post-M1-cTBS for left array: −4.34 ± 3.21 sec; pre-M1-cTBS for right array: 0.94 ± 3.05 sec, post-M1-cTBS for left array: 1.04 ± 3.12 sec; all ps > .05). The full four-way ANOVA conducted on fixation time data with factors (1) Stimulation Site (S1EYE or M1), (2) cTBS Run (pre-cTBS or post-cTBS), (3) the Side of the Array (leftward or rightward), and (4) the Half of the Array where the Fixation Fell (left or right) showed a statistically significant four-way interaction, F(1, 11) = 13.973, p = .003. A similar ANOVA performed on the number of fixations showed no statistically significant main effects or interactions (all ps > .12).
In contrast to the fixation time, the amplitude and frequency of the saccades showed no statistically significant interaction between Stimulation Site × cTBS Run × Array Side × Saccade Direction (left or right), F(1, 11) = 0.048, p > .83 for amplitude and F(1, 11) = 0.097, p > .76 for frequency. No other main effects or interactions were statistically significant.
Experiment 3 showed a shift in mean eye position and longer fixation time further from the orbit midline in opposite direction to the shift in visual sensitivity (Experiment 2). This did not support the hypothesis that S1EYE-cTBS causes a spatial bias that generalizes from visual sensitivity to visual exploration.
cTBS Effects on Calibration
Neither calibration accuracy nor precision were affected by cTBS. The mean calibration error was below 0.01° for either experiment (1, 2, or 3), cTBS run (pre- or post-cTBS) or stimulation site (M1 or S1EYE). The standard deviation of this error was below 2.56°. For each of the three experiments, paired-samples t tests showed no statistically significant difference between pre- and post-cTBS for either stimulation site (all ps > .11).
Experiment 1 showed that S1EYE-cTBS affected the amplitude of the saccades from a lateral, left or right eye rotation to a central auditory, but not visual target. We argue that the size of the saccade to auditory targets reflects the perceived eye position at the onset of the saccade. Our argument is based on the preferential coding of both auditory and visual saccade targets in eye-centered coordinates at the level of the superior colliculus (Jay & Sparks, 1984) and FEF (Russo & Bruce, 1994). Because the brain extracts the location of the auditory stimuli relative to the ears and the head, saccades to auditory targets critically depend on information of eye position to transform head- into eye-centered representations.
These changes in the perceived eye position were triggered from the same brain area (S1EYE) where we have previously demonstrated an afferent projection from the eye muscles (Balslev & Miall, 2008). Neuronavigation has placed this area in the left postcentral gyrus (Balslev, Siebner, et al., 2012). For both leftward and rightward eye position at fixation, S1EYE-cTBS caused an undershoot of the auditory-guided saccades. This suggests an underestimation of the rotation of the eye in the orbit at fixation. This effect can be predicted under two assumptions. First, we assumed that the human eye proprioceptive area S1EYE (Balslev & Miall, 2008) is organized like the proprioceptive area 3a identified in the macaque. In the macaque area 3a, the frequency of neural firing increases with the rotation of the eye in the orbit for all gaze directions (Wang et al., 2007). If the eye proprioceptive area in the human brain codes the angle of gaze in the same way, a decrease of excitability of this neuronal population would decrease in neuronal firing for a given angle of eye rotation and cause an underestimation of this angle. Second, we assumed that cTBS over S1EYE causes a decrease of excitability of this area. cTBS has been shown to decrease cortical excitability of the area underneath the coil (Huang et al., 2005). In particular, when applied at 2 cm posterior to M1, cTBS decreases the amplitude of the somatosensory-evoked potentials (Ishikawa et al., 2007), which is a correlate of the cortical excitability in the somatosensory cortex.
The absence of a statistically significant effect on saccades to visual targets on the initial position at fixation as well as on other measures of saccade kinematics (velocity and time until onset) rules out a general effect of S1EYE-cTBS on the ability to fixate or to execute saccades.
In Experiment 2, we have used S1EYE-cTBS as a method to dissociate perceived gaze direction from actual gaze direction to investigate its association with a spatially congruent shift in visual sensitivity. In line with this hypothesis, we found that S1EYE-cTBS indeed caused a decrease in the RTs for discriminating a letter when the letter was presented nearer the orbit midline. The increase in visual sensitivity toward orbit midline observed here in the RT measures was consistent with that found in visual accuracy (Balslev et al., 2011) as well as with the variation in the extrastriate cortex neural activity (Balslev, Siebner, et al., 2012) after 1-Hz rTMS in S1EYE.
Experiment 3 did not support the hypothesis that S1EYE-cTBS cause a lateral bias in visual exploration in the same direction as the increase in visual sensitivity. When participants searched a lateral array, S1EYE-cTBS shifted mean eye position away from the orbit midline, in opposite direction as the increase in visual sensitivity. This shift in mean eye position was caused by an increased fixation time in the half of the array that was further from the orbit midline. One could attempt to explain the direction of this shift in two ways. First, stimulus visibility is a well-established factor that influences oculomotor behavior during visual search (Najemnik & Geisler, 2005; Näsänen, Ojanpää, & Kojo, 2001; Jacobs & O'Regan, 1987). Humans adapt their saccades and fixations to the visibility of the array to optimize the time needed to find the target (Najemnik & Geisler, 2005). After S1EYE-cTBS, participants have lower visual sensitivity for the area of the array located further from the orbit midline. Fixating toward the outer regions of the array would bring the area of low visibility closer to the fovea. An alternative to this explanation is that the shift in mean eye position during visual search away from the orbit midline does not reflect the asymmetric visual sensitivity, but rather the shift in perceived gaze position. For instance, the participants may have perceived their gaze as being closer to the center of the array than it actually was. It is currently not known whether the oculomotor behavior during visual search is directly affected by the eye proprioceptive information. We believe this explanation is less likely, given that S1EYE-cTBS did not interfere with the participants' ability to fixate or to execute saccades under visual guidance (Experiment 1).
The results of the first two experiments show that the lateral shift in visual sensitivity induced by changing the excitability of S1EYE is spatially congruent with the rTMS-induced shift in perceived gaze direction. This spatial congruence lends support to the hypothesis that the eye proprioceptive area in the somatosensory cortex plays a role in visuospatial attention. S1EYE-cTBS increased visual sensitivity in the display area located nearer the orbit midline (Experiment 2) and shifted mean eye position during visual exploration in the opposite direction toward the area of the array located further from the body midline (Experiment 3). The spatially incongruent effects on visual sensitivity and visual exploration do not support an interpretation of the role of S1EYE in terms of visual orienting (a bias in both visual sensitivity and oculomotor behavior in the same direction). The null result when we tested for a lateral gradient in the amplitude or frequency of the incidental saccades in Experiment 2 is in line with this interpretation.
Visual sensitivity increased nearer to the orbit midline compared with further from orbit midline, although stimuli were presented at equal retinal eccentricity (Experiment 2). We argue that this difference in visual sensitivity reflects the effect of S1EYE-cTBS on visuospatial attention. Spatial attention can be defined as the selection of a location for privileged stimulus processing (Petersen & Posner, 2012; Bisley & Goldberg, 2010). The change in visual sensitivity for retinally identical stimuli has previously been interpreted as a signature of the neural processes that implement this function (Ruff et al., 2006; Cavanaugh & Wurtz, 2004; Moore & Fallah, 2004; Hilgetag & Pascual-Leone, 2001). To the best of our knowledge, spatial attention is implemented by a network that, at cortical level, consists of the FEF, areas around the intraparietal sulcus, the TPJ, and the ventral frontal cortex (Petersen & Posner, 2012; Bisley & Goldberg, 2010; Corbetta & Shulman, 2002). The current results point to the somatosensory cortex as an additional node in this network.
Because the rTMS-induced electric field can spread both locally to the adjacent areas as well as remotely, through neural connections, one could object that the effects on visual sensitivity reported here do not arise from the area underneath the coil but rather from the parieto-premotor areas, which are located near the somatosensory cortex and are likely to be interconnected with it. Against this interpretation speaks the difference between the current results and the findings of previous studies, in which TMS was applied directly over the posterior parietal cortex or the FEF. For instance, inhibitory rTMS applied at the posterior parietal foci P3/P4 can enhance the perception of stimuli located in the visual hemifield ipsilateral to the hemisphere where rTMS was applied (Hilgetag & Pascual-Leone, 2001) or bias visual exploration (i.e., fixation time) toward the hemispace located contralateral to the rTMS site (Nyffeler et al., 2008). This effect differs from the rTMS effects observed here. After S1EYE-cTBS, the RT for detecting a target as well as the mean gaze position during visual search did not show a simple dependence on the visual hemifield or the body/head hemispace but rather an interaction between these two factors. Namely, after S1EYE-cTBS a right hemifield target was detected more accurately than a left one only when both targets were presented to the left of the body/head midline, whereas to the right of the body/head midline the results were reversed. Applied over the FEF, TMS can also affect visual sensitivity. However, the variation in visual sensitivity depends on the eccentricity of the stimulus on the retina, for example, increased in the periphery compared with the center (Ruff et al., 2006). This again differs from the results reported here. Therefore, we argue that the variation in visual sensitivity after S1EYE-cTBS is unlikely to have been caused merely by a spread of the induced electric field to these areas, whose function in visuospatial attention is well established.
The current study demonstrated that visual sensitivity increases in the direction of the perceived gaze. We suggest therefore that the cortical eye position signal in the somatosensory cortex and visuospatial attention are functionally coupled. The resolution of the TMS method does not allow to conclude whether the neural population responsible for the change in visual sensitivity colocates with that coding the direction of gaze. However, this speculation seems to us reasonable, given that the changes in visuospatial attention were elicited from the same TMS coil placement that targets the human eye proprioceptive area (Balslev & Miall, 2008). We anticipate therefore that this finding will fuel further research into the role of the recently discovered eye proprioceptive area in visuospatial attention by exploring its anatomical connectivity and its functional interdependence with other nodes of the spatial attention network.
This work was supported by the Danish Medical Research Councils (09-072209) and a Marie Curie Intra-European Fellowship within the 7th European Community Framework Programme (D. B.). We thank Prof. H.-O. Karnath for comments on this manuscript.
Reprint requests should be sent to Daniela Balslev, Center of Neurology, Division of Neuropsychology, University of Tübingen, Hoppe-Seyler-Str. 3, 72076 Tübingen, Germany, or via e-mail: firstname.lastname@example.org.