Effects of inhibitory theta burst TMS to different brain sites involved in visuospatial attention – a combined neuronavigated cTBS and behavioural study

Abstract

Purpose and methods: The study sought to alter visual spatial attention in young healthy subjects by a neuronavigated inhibitory rTMS protocol (cTBS-600) to right brain areas thought to be involved in visual attentional processes, i.e. the temporoparietal junction (TPJ) and the posterior middle frontal gyrus (pMFG), and to test the reversibility of effects by an additional consecutive cTBS to the homologue left brain cortical areas.

Results: Healthy subjects showed a leftward bias of the egocentric perspective for both visual-perceptive and visual-exploratory tasks specifically for items presented in the left hemifield. cTBS to the right TPJ, and less systematically to the right pMFG reduced this bias for visuo-spatial and exploratory visuo-motor behaviour. Further, a consecutive cTBS to the left TPJ changed the bias again towards the left for a visual-perceptive task.

Conclusions: The evidence supports the notion of an involvement of the right TPJ (and pMFG) in spatial visual attention. The observations further indicate that inhibitory non-invasive brain stimulation (cTBS) to the left TPJ has a potential for reversing a rightward bias of spatial attention when the right TPJ is dysfunctional. Accordingly, the findings could have implications for therapeutic rTMS development for right brain damaged patients with visual neglect.

Keywords

Neglect cortex transcranial magnetic stimulation spatial attention exploration egocentric allocentric temporoparietal junction TPJ middle frontal gyrus MFG

1 Introduction

Unilateral spatial neglect is a frequently encountered clinical syndrome in brain-damaged people characterised by (a) an inability to attend to and report stimuli on the side opposite the lesion (contralesional) despite normal visual perception, and (b) a spatial bias for directing actions toward the hemi-space or hemi-body on the same side as the lesion (ipsilesional) (Corbetta and Shulman, 2011).

A neglect syndrome can be observed in 25– 30% of all stroke patients (Appelros et al., 2002; Buxbaum et al., 2004). In 40% of patients, neglect does not recover after one year and becomes chronic (Nijboer et al., 2013). Functional outcome of stroke patients suffering from neglect is worse than that of stroke patientswithout neglect (Nijboer et al., 2013; Nys et al., 2005), and motor recovery patterns are slower and more attenuated (Nijboer et al., 2014). Symptoms of neglect had been treated with different treatments such as visual scan training, limb activation, mental imagery training, sensory stimulation, and prism adaptation. The effectiveness of these treatments remains, however, largely unproven (Bowen et al., 2013). Non-invasive brain stimulation (NIBS) that modulates excitability and thereby function of cortical networks could become a new option to treat dysfunctional syndromes such as neglect after stroke (Nyffeler et al., 2009; Cazzoli et al., 2012; Kim et al., 2013). Yet, valid neurophysiological models for neglect treatment with NIBS are still to be established.

Spatial deficits can be separated based on the reference frame in which stimuli are coded (Marsh and Hillis 2008): neglect is most frequently egocentric (viewer-centered), with left and right hemi-spaces based on the observer’s midline; neglect can, however, also be allocentric, where the midline is defined from the central axis of a stimulus, irrespective of its position (and orientation) in the environment (object-centered).

The egocentric deficits can both be observed (i) with simple visuo-perceptive tasks, e.g. when judging whether a line is bisected in the middle, toward the left or right, and (ii) visuo-explorative tasks, e.g. when supposed to detect a number of visual targets across the visual field that are embedded in an array of distractor items.

A factorial analysis of test results in 80 right brain damaged patients (Verdon et al., 2010) supported this distinction and revealed three main factors explaining 82% of the total variance across all neglect tests, which suggested distinct components related to (1) perceptive/visuo-spatial, (2) exploratory/visuo-motor, and (3) allocentric/object-centred aspects of spatial neglect. The anatomical voxel-based lesion-symptom mapping analysis pointed to specific neural correlates for each of these components, including the right inferior parietal lobule (temporo-parietal junction, TPJ) near the supramarginal gyrus (BA 40) for the perceptive/visuo-spatial component, the right frontal cortex including the posterior part of the middle frontal cortex (BA6) (pMFG) and the inferior frontal (B45) and more anterior dorsolateral prefrontal cortex (BA46, BA10) for the exploratory/visuo-motor component, and deep temporal lobe regions for the allocentric/object-centred component. Similarly, Shirani et al. (2009) studied 137 patients within 24 hours of stroke onset with MR diffusion- and perfusion-weighted imaging and a test of hemispatial neglect that distinguishes between viewer-centered and stimulus-centered neglect. Using multivariable linear regression, severity of hypoperfusion in the angular gyrus was the only variable that significantly and independently contributed to severity of viewer-centered neglect, severity of hypoperfusion of superior temporal cortex was the only variable that independently and significantly contributed to severity of stimulus-centered neglect.

When spatial attention has been investigated in healthy subjects with functional imaging techniques, the corresponding topographic maps have, however, not been reported in these ventral regions typically damaged in neglect patients, consistent with the lack of evidence for their involvement in spatial attention; conversely, dorsal frontoparietal regions that contain these signals, i.e. process spatial attention, are not typically damaged in strokes that cause neglect (Corbetta and Shulman, 2011). To resolve this paradox, these authors hypothesize that damage to right ventral frontoparietal cortex in neglect patients (including TPJ) does impair nonspatial functions, but consecutively hypoactivates the right hemisphere and thereby causes less activation of the dorsal attention network thought to process spatial attention; the ventral-dorsal interactions would thus indirectly cause the egocentric spatial bias that is the hallmark of the neglect syndrome. Supporting this assumption, Corbetta et al. (2005) showed in a fMRI study in patients suffering from neglect after right hemispheric lesions that lesions in the ventral attentional system, excluding the dorsal parietal cortex, trigger functional changes in ipsi- and contralesional anatomically intact areas. Whereas hypoactivity was observed in the ipsilesional dorsal parietal cortex, relative hyperactivity was found in the contralesional dorsal parietal cortex. Furthermore, the relative hyperactivity in the left dorsal parietal cortex correlated with the number of neglected targets in the left hemispace. Finally, neglect recovery was accompanied by restoration of the interhemispheric balance in the dorsal parietal cortex.

Accordingly, interhemispheric imbalance has been discussed as a pathophysiological mechanism in visuospatial neglect. It is suggested that after a lesion of the right hemisphere the mutual transcallosal inhibition is impaired, resulting in an increased activity of the left hemisphere.

So far, rTMS effects on spatial attention in healthy subjects and stroke patients with neglect had been tested for parts of the dorsal attention network, e.g. the posterior parietal cortex (PPC), where inter-hemispheric rivalry dynamics could be observed. Cazzoli et al. (2009) investigated the interhemispheric balance of attention in healthy subjects by using a free visual exploration task and by interfering with the neural activity of the posterior parietal cortex (PPC) of either hemisphere using an inhibitory transcranial magnetic stimulation routine with theta burst stimulation (TBS). Subjects explored colour photographs of real-life scenes presented on a computer screen under four conditions: (i) without TBS; (ii) after TBS over the right PPC; (iii) after TBS over the left PPC; and (iv) after TBS over the right PPC and, after the first half of the task, over the left PPC. Eye movements were measured, and distribution of mean cumulative fixation duration over screen halves was analysed. TBS over the right PPC resulted in a significant rightward shift of mean cumulative fixation duration for 30 min. The shift could be reversed when a subsequent train of TBS was applied over the left PPC. Nyffeler et al (2009) aimed to test in stroke patients whether parietal TBS over the unaffected hemisphere can induce an improvement of visual neglect by reducing the interhemispheric inhibition. In their study, eleven patients with left-sided visual neglect attributable to right hemispheric stroke were tested in a visual perception task. The application of 4 TBS trains over the left PPC significantly increased the number of perceived left targets up to 32 hours. Thus, the results indicate – at least for the contralesional posterior parietal cortex – that rTMS can ameliorate neglect symptoms in stroke patients with neglect.

The current sham-controlled rTMS experiment was motivated by this background, yet targeted areas of the ventral attentional system. Specifically, the study addressed whether spatial visual processes, i.e. perceptive/visuo-spatial, exploratory/visuo-motor, and/or allocentric/object-centred aspects could experimentally be altered in healthy subjects by an inhibitory rTMS (cTBS) (a) to the right TPJ, frequently damaged in neglect patients showing perceptive/visuo-spatial deficits, or (b) the right pMFG, belonging to a network that when damaged was associated with exploratory/visuo-motor deficits (Verdon et al., 2010).

Noteworthy, both stimulated areas, i.e. right TPJ and pMFG are considered as parts of the ventral attention network typically damaged in neglect patients, and yet in studies with healthy subjects there has been a lack of evidence for their involvement in spatial attention (Corbetta and Shulman, 2011). As opposed to human lesion studies with complex brain pathologies, in the current experiment these cortical regions and their contribution to the mentioned aspects of spatial attention could selectively be probed experimentally in healthy subjects. Assuming that spatial visual processes could be altered, i.e. spatial bias induced or changed by cTBS of these right brain areas, it was further investigated whether a subsequent inhibitory rTMS (cTBS) of the homologue area in the left hemisphere could reverse the observed behavioural deficits. If that was possible, interhemispheric imbalance as a pathophysiological mechanism for spatial attention deficits could be entertained not only for areas of the dorsal, but also of the ventral attentional network. Thereby, the involvement of these areas in spatial visual processing and any effects of targeted inhibitory rTMS as well as their “therapeutic” reversal was sought to be investigated.

2 Methods

The research question was addressed with a sham-controlled cohort study with healthy subjects who were investigated behaviourally (visuospatial/neglect tests) before and shortly after either inhibitory cTBS or sham stimulation of pre-defined cortical areas of interest using a within-subject analyses of effects. Prior to commencing the study, its protocol had been approved by the local ethics committee of the Ernst-Moritz-Arndt-Universität Greifswald.

2.1 Subjects

Candidate subjects were given information about the study verbally and in writing and were able to participate if they fulfilled inclusion (healthy, age 18 to 30, right-handed), did not match exclusion criteria, and gave written informed consent. An increased health risk with rTMS (TBS) is considered for people with incorporated ferromagnetic devices, during pregnancy, as well as with a history of epilepsy or an epileptic seizure; subjects with such a history were excluded.

Ten healthy right-handed adults (students) aged 25 to 30 (mean 27.3, s.d. 3.4 yrs.) were locally recruited, 7 are females, 3 males.

2.2 Experimental procedure

The first experimental interest was to assess any alterations (left- or rightward bias) of spatial attention when cTBS was applied to one of two right hemisphere target sites, i.e. the temporoparietal junction, TPJ, or the posterior middle frontal gyrus, pMFG. To account for any non-specific influences a sham stimulation of right hemisphere primary motor cortex (M1) was also used for comparison.

A specifically linked experimental interest was then to test whether an additional cTBS to the corresponding homologue area of the left hemisphere applied during the same session could reverse any chances of spatial attention induced by the first cTBS. For the second stimulation either cTBS or sham stimulation had been used to account for “placebo”-effects.

Since effects of single cTBS sessions are likely to vanish within 24 hours (Nyfeller et al., 2009), experimental sessions were spaced with at least 2 days intervals for individual subjects. Specifically, 5 cTBS sessions were planned in 2 consecutive weeks, on Monday, Wednesday, Friday. To account for any remaining cross-over or test repetition effects, sequence of stimulation sites (stimulation combinations 1– 5, see below) was counterbalanced across participants.

Structure of each TBS session (duration appr. 1 hour):

coregistration of head and coil (neuronavigation),

baseline assessment (behavioural tests, appr. 10 minutes),

cTBS or sham to one of three right hemisphere targets (first intervention),

followed by 8 minutes rest,

post cTBS assessment 1 (behavioural tests, appr. 10 minutes),

cTBS or sham to left homologue hemisphere target (second intervention),

followed by another 8 minutes rest,

post cTBS assessment 2 (behavioural tests, appr. 10 minutes).

The sequence of events is presented in Fig. 1.

Combinations of cTBS were:

“stimulation 1”: cTBS600-R_TPJ first, sham-L_TPJ second,

“stimulation 2”: cTBS600-R_TPJ first, cTBS600-L_TPJ second,

“stimulation 3”: cTBS600-R_pMFG first, sham-L_pMFG second,

“stimulation 4”: cTBS600-R_pMFG first, cTBS600-L_pMFG second,

“stimulation 5”: sham-R_M1 first, sham-L_M1 second.

Each combination (stimulation 1 to 5) was applied in a different session at least two days apart.

To prevent that any test repetition effects (different performance levels at baseline across sessions) could confound the analyses of cTBS vs. sham effects, we pseudorandomly and systematically changed the sequence of stimulation combinations (1– 5) across subjects over the experimental days.

Note: While the order of these stimulation combinations was pseudorandomised across subjects, for the statistical analyses and the presentation of results, the combinations were given the above mentioned descriptors, i.e. stimulation1 to 5.

2.3 Neuronavigated transcranial magnetic stimulation

Anatomical brain measurements of all participants were obtained using magnetic resonance imaging (MRI). An anatomical T1-weighted three-dimensional Magnetization Prepared Rapid Gradient Echo (MPRAGE) image was acquired for each subject at a 3T Siemens Magnetom Verio (Siemens, Erlangen, Germany) equipped with a 32-channel head coil. The total number of sagittal anatomical slices amounted to 176 (TR = 1900 ms, TE = 2.52 ms, α= 90°, voxel size = 1×1×1 mm 3). We performed a surface reconstruction to recover the spatial structure of the cortical sheet based on the white-grey-matter boundary using BrainVoyagertrademark TMS neuronavigation software (TMS Neuronavigator edition of BrainVoyager QX 2.1 by Brain Innovation B.V., Maastricht, NL).

We then identified the target brain regions, namely the temporoparietal junction (TPJ) and the posterior middle frontal gyrus (pMFG) on either side of the brain (compare Fig. 2). The pMFG stimulation site was identified as lying in the middle (dorso-ventral) of the gyrus in the posterior part of the MFG; the TPJ stimulation site was determined as the cortical area that lies in the temporo-parietal junction just above a line that lies over the temporal sulcus, and again in the (cranio-caudal) middle of the selected gyrus

Motor evoked potentials (MEPs) were used to determine the coil position over M1 (“hot spot”) that evokes the best response in the left abductor pollicis brevis muscle (APB). The location of this spot (right M1) and its left brain homologue was also marked using BrainVoyager trademark.

Each participant’s anatomical MRI and head and brain surface models were used for stereotactic co-registration of the participant’s brain with the TMS coil. This enabled online control and re-test-reliability of coil positioning during each session and across days.

Subjects were seated in a reclining chair and instructed to remain relaxed throughout the application of TMS. During MEP assessment, surface electromyography (EMG) from participants’ APB was monitored using the motor evoked potential unit of (Dantec Keypoint^® by Alpine Biomed ApS, Skovlunde, DK). Application of TMS was performed with a 75 mm figure-of-eight passively cooled coil (MCF-B65 or corresponding placebo coil; placebo coil: the coil’s magnetic shield provides a field reduction of approximately 90% ; it has an identical mechanical and optical outline and sound level as the MCF-B65) and the MagPro X100 Magnetic Stimulator (MagVenture A/S, Farum, DK). The TMS coil (MCF-B65 or placebo) was oriented tangentially to the scalp with the handle pointing back and away from midline at 45° for all stimulation sites.

The TBS protocol used short bursts (3 stimuli) of 50 Hz rTMS which were repeated at a rate in the theta range (5 Hz). The rTMS sessions consisted of a continuous 40 second train of TBS with 600 stimuli (cTBS-600) (Huang et al., 2005). Applied continuously (cTBS), the net effect of TBS is inhibitory; cTBS-600 can temporarily suppress local cortical excitability for about 60 minutes. In this experiment, an intensity that equals 80% of the active motor threshold was applied; the AMT being the intensity that evokes an MEP of > = 200μV in > = 5 out of ten trials while the subjects perform an isometric contraction at a level of 20% of her/his maximum voluntary contraction. Note that because the bursts are given at high frequency, the intensity of stimulation should be quite low, and certainly below resting motor threshold (Rossi et al., 2009).

2.4 Assessment of spatial attention – outcome measures

Assessment of spatial attention with the below mentioned tests was performed immediately before (time 1) and (starting at 8 minutes) after each cTBS and/or sham combination (time 2 or time 3, respectively) on each stimulation day (pre and post tests).

The sequence of assessment was standardised as follows:

Bells cancellation,

Line bisection – active motor, paper and pencil version

Line bisection – perceptual PC version (visual discrimination)

Ota search task - perceptual PC version.

Assessments and stimulation were performed in the afternoon.

2.4.1 Bells cancellation (Gauthier et al., 1989)

Subjects were asked to mark all bells disseminated among distractors (i.e. other symbols), on an A4 sheet of paper presented horizontally. The main score was the difference (BEOD = BELO – BERO) between omissions on the left side (BELO) relative to the right side of the sheet of paper (BERO).

While this test had been designed for stroke patients without time limit, a time limit was introduced for the study to increase task difficulty and thereby make the observation of omissions and experimentally induced behavioural changes among healthy subjects more likely. For this purpose, the time limit was set to 80% of individual time needed to fulfil the task at a pre-visit under the instruction to perform the task both swiftly and with precision. In addition, several test forms (i = 10) with the same type and number of bells and distractors, but non-identical spatial arrangement was used to prevent learning effects with test repetition (in each of 7 evenly spaced areas of the A4 paper the type and the number of symbols was the same for all test forms to match the overall spatial distribution of bells and distractors).

2.4.2 Ota search task (Ota et al., 2001)

In the original task, patients are asked to mark all circles with a gap (on the circle’s left or right side) among other circles without gap. Two scores are obtained to reflect egocentric neglect (omissions of targets on the left side of the sheet) and allocentric neglect (omission of targets with a gap on their left).

For this investigation assessing healthy subjects a computerised test was created that uses comparable items and addresses similar constructs: An array of 30 black circles was permanently shown distributed across a screen. Dynamically, a fixation cross was shown in the middle of the screen for 700 ms followed by the brief presentation (500 ms) of one of the circles in red as either full circle, or with a gap either on its left or its right side. The subjects were asked to press a L, M, or R key depending on whether a gap on the left side (L), no gap (M), or a gap on the right side (R) characterised a single highlighted (red) item. Both reaction time and correctness of responses were documented. Once a response has been generated by the subject the fixation cross was again presented followed by another item. Reaction times (RT) and error counts (E) were averaged (A) separately for items on the far right (OR2RT; OR2E), right (OR1RT, OR1E), or left (OL1RT, OL1E), or the far left (OL2RT, OL2E) 25% of the screen (egocentric factor), and (B) for items with a left-sided gap (OGLT, OGLE), or right-sided gap (OGRT, OGRE), or no gap (ONGT, ONGE), respectively (allocentric factor). Neglected items were thought to reflect egocentric errors; allocentric errors were assumed when items had been misclassified (as “closed” when “open”, or less likely “open” when “closed”).

2.4.3 Line bisection (Schenkenberg et al., 1980)

Subjects were asked to mark the middle of five 20 cm horizontal lines, presented individually on an A4 sheet of paper. The score was the magnitude of left- or rightward deviation from the true centre (in millimetres); results for the five lines were averaged (LB).

In addition to the paper and pencil version, a computerised timed visual discrimination version of the line bisection task (similar to Loetscher et al., 2012) was used as a task that was potentially more sensitive to cTBS effects. Subjects were presented 20 cm long horizontal lines that were bisected either in the middle (0 mm; i = 3), towards their left (– 1, – 2, – 3, or – 4 mm; i = 12 [3 lines of each type]), or towards their right (+1, +2, +3 or +4 mm; i = 12 [3 lines of each type]). In total, 27 black lines with a length of 20 cm were presented in a pseudo-randomized order either on the left (i = 9), centre (i = 9) or right side (i = 9) of the screen. There were 9 lines per side of presentation with the lines in the lateral conditions being shifted 4 cm to the left and right, respectively (without vertical shift). Subjects were presented each line for one second, two seconds after a cue (dot at midscreen above the level of the lines) and had then to decide whether the presented lines had been divided either towards their left or right or in the middle by pressing one of three touch switches (L, M, R). Reaction times (RT) were averaged for items presented in the middle of the screen (LMRT), and either on the right (LRRT) or on the left (LLRT) (field factor). A performance score summarised the quality of performance (LP) and was calculated as average of all values attributed to individual responses (that could range from – 5 to +5): For items with lines bisected in the middle, responses that judged them as bisected in the middle, or towards the left or right were given a score of 0, – 1, or +1, respectively; for items with lines bisected 1 mm towards the left, responses that judge them as bisected towards the left, middle or right were given a score of 0, +1, or +2; for items with lines bisected 2 mm towards the left, responses that judge them bisected towards the left, middle or right were given a score of 0, +2, or +3; for items with lines bisected 3 mm towards the left, responses that judge them bisected towards the left, middle or right were given a score of 0, +3, or +4; for items with lines bisected 4 mm towards the left, responses that judge them bisected towards the left, middle or right were given a score of 0, +4, or +5; a corresponding scoring was used for lines bisected towards the right, e.g. for items with lines bisected 3 mm towards the right, responses that judge them bisected towards the left, middle or right were given a score of – 4, – 3, or 0. The scoring system guaranteed that the more deviant the judgment, the higher a score. This performance score was calculated (a.) as sum for all stimuli (object-centred evaluation) and (b.) separately for lines presented either in the middle of the screen (LMLP), or on the right (LRLP) or on the left (LLLP) side of the screen (evaluation of field effect).

Taken together these tests are thought to assess

a component reflecting bias in exploratory visuo-motor behaviour (left- or right-sided misses in the Bells cancellation test and Ota search task),

perceptive visuo-spatial bias (deviation on line bisection), as well as

allocentric visuo-spatial behavioural differences, indicated by errors or an increased reaction time that can specifically be observed for circles with left-sided (or right-sided) gaps in the Ota task.

2.5 Statistical analyses

2.5.1 Sample size determination

Assuming an experimentally induced effect size of 0.5, selecting an alpha error probability of 0.05, a power of 0.90, and planning 5 measurements in a repeated measures ANOVA, a total sample size of 8 subjects was necessary to corroborate effects statistically (estimated with GPower 3.1 software; Faul et al., 2009). A total number of 10 subjects was intended to be recruited to compensate for up to 2 (20%) drop-outs.

2.5.2 Analysis of outcome measures

General linear models within a repeated-measures ANOVA design were used. Within a repeated-measures ANOVA we assessed the effect of the type of cTBS (factor “stimulation”, i.e. 5 combinations as introduced in section ‘2.2. Experimental procedure’) on outcome measures, e.g. the behavioural measures at baseline, after the first, and after the 2nd stimulation in each session (factor “time”). F values presented for these models are partial F values (based on type III sums of squares). The Huynh-Feldt correction for the degrees-of-freedom was used where applicable and the value of ɛ is reported while the adjusted p level is labelled “H– F.”

The repeated measure ANOVA was augmented by embedded, pre-planned univariate secondary analyses. With regard to the type of the two combined stimulations per experimental session (cTBS + contralateral sham or cTBS + contralateral cTBS to either TPJ or pMFG) (factor ‘stimulation’), any effects of these stimulations were compared to the sham only stimulation condition (sham + sham to M1). With regard to the factor ‘time’ (i.e. within session changes) any within session effects were analysed by planned comparisons between the 2nd (‘time 2’) or 3rd assessment (‘time 3’) within each session with the first (baseline) assessment (‘time 1’) of that session; thereby, differences to the status before any stimulation (on that day) were analysed statistically. The analyses of interactions between factors ‘stimulation’ and ‘time’ had been analysed based on these pre-specified contrasts (only). Thus, while mean and s.d. are presented for the specified observations in the text and figures, all statistical analyses were based on these pre-planned comparisons (i.e. intra-individual contrast to the sham only condition and/or baseline data).

3 Results

3.1 Line bisection - paper version

Subjects were asked to mark the middle of five 20 cm long horizontal lines, each presented individually on an A4 sheet of paper. The score was the magnitude of left- or rightward deviation from the true centre (in millimetres). Results for the five lines were averaged (score ‘LB’).

The statistical analysis revealed a main effect for the factor ‘time’, i.e. changes across assessments within the sessions (F(2,18) = 5.25, p = 0.0181):

A minor leftward bias at the beginning of the sessions was reduced with repeated testing within sessions (LB, mean/s.d.: time 1 – 1.31/2.44 mm; time 2 – 0.61/2.45 mm; time 3 – 0.80/2.35 mm). Since no interaction with the type of stimulation was seen, an effect of stimulation could not be corroborated. The changes rather reflect a within session test repetition effect.

3.2 Line bisection - PC version

Single bisected lines were presented for one second each, two seconds after a cue (dot at midscreen above the level of the lines). The subjects had then to perceive and decide whether the presented line had been divided either towards its left (L), or right (R), or in the middle (M) by pressing one of three touch switches (L, R, or M, respectively).

A performance score summarised the performance quality with this test and was calculated as average of values attributed to individual responses (ranging from – 5 to +5), 0 denoting perfect performance. This performance score was calculated both as sum for all stimuli (line bisection performance – ‘LP’) (object-centred evaluation), and separately for lines presented either in the middle of the screen (‘LMLP’), or on the right (‘LRLP’), or on the left (‘LLLP’) side of the screen (evaluation of a field effect).

Similarly, reaction times to these stimuli were averaged for items presented in the middle of the screen (‘LMRT’), or on the right (‘LRRT’), or on the left (‘LLRT’) side of the screen, respectively (field effect on reaction time).

For the univariate contrast analyses of any field effects, scores for stimuli presented on either the left or right side of the screen were compared to scores obtained for items presented in the middle of the screen.

Analyses independent of the field of presentation

When performance was analysed independent of the field of presentation (left, middle, or right), no systematic effects (changes within sessions, after stimulation) were found.

Effects for the field of presentation

There were performance changes after non-invasive brain stimulation when the bisected lines were presented in the left field:

At baseline, judgement for lines presented on the left side of the screen was as if the bisection was more towards the right of each line than it actually was(indicating a leftward bias of perception, i.e. a subjective middle that was left of the objective middle of the lines). This decreased when the right TPJ had received the inhibitory stimulation, i.e. a perception bias towards the left was reduced after cTBS to the right TPJ (left field, performance; cTBS_R_TPJ & sham, mean/sd: time 1 0.31/0.52; time 2 0.17/0.46; time 3 0.20/0.59; left field×stimulation1×time 2: F(1,9) = 7.15,p = 0.0254) (compare Fig. 3A). The effect showed a tendency to be reversed, when cTBS to the right TPJ was followed by cTBS to the left TPJ (left field, performance; cTBS_R_TPJ & cTBS_L_TPJ, mean/sd: time 1 0.22/0.55; time 2 0.20/0.57; time 3 0.29/0.56; left field×stimulation2×time 2: F(1,9) = 9.26, p =0.0140; left field×stimulation2×time 3: F(1,9) =4.75, p = 0.0572) (compare Fig. 3B).

A similar reduction of the leftward bias of perception was seen when the right pMFG was stimulated; numerically (but n.s.) this was reversed after cTBS of the left pMFG (left field, performance; cTBS_R_pMFG & cTBS_L_pMFG, mean/sd: time 1 0.29/0.54; time 2 0.17/0.55; time 3 0.40/0.51; left field×stimulation 4×time 2: F(1,9) = 11.28, p = 0.0084).

With regard to reaction times, there was a considerable field effect (F(2,18) = 102.75, p < 0.0001). Reaction times to stimuli on either the left or the right were higher than for stimuli presented in the middle of the screen (reaction time, mean/sd: left 737/80 ms, middle 665/80 ms, right 748/79; left vs. middle F(1,9) = 89.20, p < 0.0001, right vs. middle F(1,9) = 190.27, p < 0.0001).

After cTBS stimulation of the right TPJ reaction time increased for stimuli presented in the left field; the effect was no longer present after consecutive cTBS of the left TPJ (left field, reaction time; cTBS_R_TPJ & cTBS_L_TPJ, mean/sd: time 1 724/72 ms, time 2 740/61 ms, time 3 725/73; left field×stimulation2× time 2 F(1,9) = 10.52, p = 0.0101) (compare Fig. 4).

cTBS of the right pMFG increased reaction times (independent of the field of presentation), statistically corroborated after the consecutively following sham stimulation (reaction time; cTBS_R_pMFG & sham, mean/sd: time 1 684/65 ms, time 2 700/66 ms, time 3 717/78; stimulation3×time 3 F(1,9) = 9.09, p = 0.0146).

3.3 Bells cancellation test

Subjects were asked to mark all bells disseminated among distractors (i.e. other symbols) on an A4 sheet of paper. Since the time limit of the task was set to 80% of individual time needed to fulfil the task at a pre-visit it was clear that errors (omission of bells) would be observed. Omissions of bells on the left side of the paper (‘BELO’) and on the right side of the paper (‘BERO’) were analysed separately as well as the intra-individual difference between these error scores (‘BEOD’ = BELO – BERO).

cTBS affected the occurrence of omissions both on the right (BERO; stimulation × time: F(8,72),p(H-F) = 0.0122, ɛ= 0.72), on the left side of the paper (BELO), as well as the intra-individual differences between omission on the right and the left side (BEOD; stimulation × time; F(8,72), p(H-F) = 0.0029, ɛ= 0.93).

The intra-individual difference scores (BEOD) were on average negative indicating more omissions for bells on the right side of the paper as compared to bells on the left side of the paper (i.e. a leftward bias of attention). This left-right difference became significantly less after cTBS to the right TPJ (BEOD; cTBS_R_TPJ & sham, mean/s.d.: time 1 – 3.20/3.55; time 2 – 0.60/3.20; time 3 – 0.40/3.20; stimulation1×time 2: F(1,9) = 8.79, p = 0.0158) (compare Fig. 5).

The effect could, however, not be corroborated statistically for the 3rd assessment within these sessions (cTBS_R_TPJ & sham, stimulation1×time 3: n.s.) or a comparable situation on another day (cTBS_R_TPJ & cTBS_L_TPJ, stimulation2×time 2: n.s.).

An effect of cTBS to the right pMFG (followed by either cTBS or sham to the left MFG) on omissions with the Bells cancellation test could not be corroborated.

3.4 Ota task – PC version

3.4.1 Field effects

Errors were averaged for the far left (‘OL2E’), left (‘OL1E’), right (‘OR1E’), and the far right (‘OR2E’) 25% of the screen, respectively (egocentric factor). For the univariate analyses, these averages were statistically compared with the intra-individual mean across these measures.

Errors when processing stimuli in the peripheral fields were symmetrically considerably more frequent than when reacting to stimuli presented in the more central parts of the screen on the left or right (factor field: F(3,27) = 29.94, p(H-F)<0.0001, ɛ= 0.54) (errors, mean/s.d.: OL2E 6.5/4.0; OL1E 2.1/2.1; OR1E 2.1/2.4; OR2E 6.3/4.3).

After cTBS to the right TPJ and sham to the left TPJ there was a slight increase in errors at the end of the session (errors; cTBS_R_TPJ & sham_L_TPJ, mean/s.d.: time 1 4.6/4.2, time 2 4.3/3.9 (n.s.), time 3 4.8/4.0, stimulation1×time 3 F(1,9) = 11.32, p = 0.0083).

Further, when cTBS was applied to the right pMFG (followed by sham to the left pMFG) errors increased in the far left part of the field while they decreased somewhat in the right field close to midline at the end of the session (OL2E, cTBS_R_pMFG & sham_L_pMFG, mean/s.d.: time 1 5.6/3.7, time 2 6.6/4.0 (n.s.), time 3 7.1/4.0, field1× stimulation3×time 3 F(1,9) = 8.95, p = 0.0152; OR1E, cTBS_R_pMFG & sham_L_pMFG, mean/s.d.: time 1 1.8/1.8, time 2 2.4/2.8 (n.s.), time 3 1.5/2.3, field3×stimulation3×time 3 F(1,9) = 14.44, p = 0.0042).

Similarly, reaction times were averaged for the far left (OL2RT), left (OL1RT), right (OR1RT), and the far right (OR2RT) 25% of the screen, respectively (egocentric factor), and statistically compared against their intra-individual mean across these measures.

Reaction time to stimuli in the peripheral fields was symmetrically higher than stimuli presented in the more central parts to the right or left (factor field: F(3,27) = 49.54, p(H-F) <0.0001, ɛ= 0.47) (reaction time, mean/s.d.: OL2RT 692/85 ms; OL1RT 622/67 ms; OR1RT 607/69 ms; OR2RT 681/83 ms).

cTBS to the right TPJ was associated with a higher reaction time for stimuli presented in the far right field (OR2RT, mean/s.d.: stimulation1 691 ms/100 ms, stimulation5 679 ms/67 ms; field4×stimulation1 F(1,9) = 6.82; p = 0.0282; stimulation1 time1 684 ms / 100 ms, time 2 711 ms/ 121 ms, time 3 678 ms/83 ms).

An effect of cTBS to the right pMFG (followed by either cTBS or sham to the left MFG) on reaction times in this task could not be corroborated.

3.4.2 The allocentric factor

Coding of allocentric errors was based on the fact that items had been misclassified as “closed” when in fact “open” with a left-sided gap (‘OGLE’) or right-sided gap (‘OGRE’), or less likely when misclassified as “open” when “closed” (no gap, ‘ONGE’).

Statistically, with pre-planned contrasts embedded in the repeated measure ANOVA either errors for items with left-sided gaps or right-sided gaps were separately compared to errors for items with no gap.

Compared to the session with sham stimulation only, errors to stimuli with left-sided gaps were numerically higher in the sessions with cTBS, an effect that was statistically significant in the condition with cTBS to the right pMFG and then left pMFG (OGLE, errors, mean/s.d.: sham & sham 0.43/0.57; right TPJ & sham 0.97/0.93, n.s.; right TPJ & left TPJ, 0.63/0.72, n.s.; right pMFG & sham 0.93/1.23, n.s.; right pMFG & left pMFG 0.80/0.92, F(1,9) = 7.88, p = 0.0205).

Reaction times were averaged for items with a left- (OGLT) or right-sided gap (OGRT), or no gap (ONGT), respectively.

Reaction time to stimuli with gaps on either the left or right side was shorter than for stimuli without gaps (factor gap: F(2,18) = 19.37, p < 0.0001) (reaction time, mean/s.d.: OGLT 617/76 ms, F(1,9) = 30.98, p =0.0003; OGRT 636/72 ms, F(1,9) = 13.96, p = 0.0047; ONGT 680/61 ms).

In addition, reaction time to stimuli with left-sided gaps improved with re-testing (factor gap×time: F(4,36) = 3.20, p = 0.0239) (OGLT, reaction time, mean/s.d.: time 1 626/73 ms; time 2 615/76 ms, F(1,9) = 7.66, p = 0.0218; time 3 609/78 ms, F(1,9) =5.46, p = 0.0443).

No effect of cTBS to either the right TPJ or pMFG (followed by either cTBS or sham to the left TPJ or MFG) on reaction times in this task could becorroborated.

4 Discussion

4.1 Line bisection

The results of the paper and pencil line bisection task indicated that the subjective horizontal midline showed a minor bias towards the left (on average 1.3 mm at baseline of the sessions). This decreased slightly with test repetition, but was not altered by brain stimulation. This observation is congruent to meta-analytic results indicating that especially young healthy subjects show a leftward bisection error in line bisection tasks (i.e., a so called “pseudoneglect”) (Jewell and McCourt, 2000).

The effect with the paper and pencil line bisection task was mirrored by an effect when bisected lines were presented on a PC screen and subjects had to perceive and discriminate whether the lines were bisected in the middle, towards the left or towards the right. When these lines were presented in the left part of the PC screen, they were on average judged as if they had been bisected slightly more towards the right than they had actually been. This indicated some bias of the perceived middle towards the left for items (lines) presented in the left part of the visual field, i.e. the perceived middle of the line was subjectively allocated somewhat left from its true centre.

cTBS of the right TPJ reduced this effect partially, while a consecutive cTBS of the left TPJ produced a tendency to reverse the effect of the first stimulation (compare Fig. 3A and B). In addition, reaction time in response to these stimuli increased after cTBS to the right TPJ while returning to baseline after cTBS to the left TPJ (compare Fig. 4).

These observations would be in line with the assumptions that the right TPJ (a.) is specifically involved in visual perception of items in the left visual space (compare effect of cTBS on reaction time), and (b.) plays a role in producing a visual-perceptive bias towards the left for items presented in the left part of the visual space (compare effect of cTBS on performance), and (c.) that a balance between right and left TPJ and its changes induced by cTBS affect the observed visual-perceptive bias towards either side (compare effect of cTBS on performance).

Qualitatively comparable effects on performance for items presented in the left part of the visual space were also seen after cTBS to the right (and then left) pMFG. These effects could, however, only partially be corroborated statistically. A negative effect of cTBS to the right MFG on reaction time was, however, independent of the field of presentation.

Taken together, both right and left TPJ and pMFG seem to contribute to the visual perceptual task. Both right TPJ and right pMFG seem to play a role in producing a visual-perceptive bias towards the left (in the visual space) for items presented in the left visual field. cTBS to the right TPJ and right pMFG can reduce this bias, while an additional cTBS to the left TPJ can shift the visual-perceptive bias back to the left for items presented in the left part of the visualspace.

In line with these observations are results of previous TMS studies that indicated that rTMS to right parietal (P4 electrode position) and frontal areas(F4 electrode position) induced a rightward shift of attention in healthy subjects (Brighina et al., 2002; Fierro et al., 2000).

4.2 Bells test

Since the time limit of the task was set to 80% of individual time needed to fulfil the task, omissions occurred even among these healthy subjects. It was interesting to note that on average omissions, i.e. bells that were not cancelled, occurred more on the right side of the paper than on the left. This might indicate a cultural phenomenon, i.e. to start to work more on the left side of a paper (like when reading or writing), or alternatively could be an effect of a leftward bias of visual attention and exploratory visuo-motor behaviour in healthy subjects, mirroring the effects seen for visual perception.

cTBS affected the occurrence of omissions in a spatially directed way. Specifically, cTBS to the right TPJ reduced the observed leftwards bias (led to more omissions of the left and less omissions on the right) and thus decreased the left-right difference of omissions. This effect of cTBS to the right TPJ was substantial (left-right difference at baseline – 3.20 on average, after cTBS to the right TPJ – 0.60) and could not be documented for other stimulation conditions. A reversal of the effect by a consecutive cTBS to the left TPJ was not observed.

These observations would be in line with the assumptions that the right TPJ (a.) is not only involved in visual perception, but also plays a functional role in exploratory visuo-motor behaviour, and (b.) is physiologically involved in producing an exploratory visuo-motor bias towards the left (with more errors on the right).

Further, the data after cTBS denote that a dysfunctional state of the right TPJ can change a bias for visual attention and exploratory visuo-motor behaviour towards the right. Yet, there was no prove in this data set for an intervention (e.g. cTBS to the left TPJ) that could reverse this effect.

4.3 Ota task

With a PC version of the task, an array of 30 black circles was permanently shown distributed across a screen. Dynamically, a fixation cross was shown in the middle of the screen followed by the brief presentation of one of the 30 circles in red, as either full circle, or with a gap either on its left or its right side. Subjects had to react to the occurrence of a red item (field effect, egocentric factor) and decide if it had a gap on the left or right or none (allocentric factor).

Field effects

With this perceptual PC version of an Ota task, a substantial field effect was observed. Overall, more errors were made and higher reaction times were observed for items in either the far right or the far left 25% of the screen (egocentric factor).

Effects of cTBS were only minor after cTBS to the right TPJ with a small increase in errors (without field effect) and an increased reaction time for stimuli in the far right field, an effect that is not readilyexplained.

When cTBS was applied to the right pMFG, errors increased in the far left part of the field while they decreased in the right field close to midline.

Overall, there were no big effects of cTBS on this visuo-perceptual task regarding field effects. Nevertheless, the observations made would be in line with the assumption that the right pMFG is involved in visuospatial perception and that its dysfunctional state negatively affects visual stimulus detection in the left visual space, or shifts visual attention towards the right. While previously right frontal TMS (including the frontal eye file, FEF) has consistently been found to affect performance in both hemifields (unlike parietal TMS) (e.g. Duecker et al., 2013), the current data with neuronavigated inhibitory rTMS to pMFG suggests that this frontal area can have a spatially biased functional role.

The allocentric factor

Reaction time to stimuli with gaps on either the left or right side was shorter than for stimuli without gaps. This reaction time effect might be explained by the fact that the items with gaps were more different from the background items than those without gap facilitating their detection and processing.

In addition, reaction time to stimuli with left-sided gaps improved during sessions indicating a test repetition effect. The fact that such a practice effect was documented only for a subset of stimuli might indicate a specialised information processing. It is further of interest to note that this subset of stimuli was affected by cTBS. Errors for stimuli with left-sided gaps showed a tendency to be bigger in sessions with cTBS, statistically corroborated for cTBS to the right (and then left) pMFG.

Taken together the data is suggestive of a left vs. right allocentric factor to play some role in visual perception and that the pMFG is part of a network involved in this type of processing.

4.4 General discussion

The data suggest a slight leftward bias of the egocentric perspective for both visual-perceptive and visual-exploratory tasks (line bisection task, active motor; line bisection task, visual discrimination; bells test) in young healthy subjects. The data would support the assumption of a currently prevailing model that a bihemispherically organised visual attention system - where each intra-hemispheric network orients visual attention towards the other side of the space - shows a slight dominance towards the right hemisphere/left visual space (Corbetta and Shulman, 2011).

This model of spatial and non-spatial visual attention processes further assumes that a ventral attentional network in the right ventral fronto-parietal cortex (including the TPJ and pMFG) subserves of a nonspatial mechanisms for reorienting, detection, and arousal.

The effects of an inhibitory non-invasive brain stimulation (cTBS) of the right TPJ in the current experiment would, however, be in line with the assumption that this area could directly be involved in mechanisms controlling spatial attention towards the left visual space, both for visual-perceptive (visual discrimination of bisected lines) and visual-exploratory tasks (bells test). E.g. cTBS to the right TPJ reduced a substantial right-left bias for omissions with the bells test.

In line with the presented data, several previous TMS studies have shown that disruption of the ventral network can indeed influence attentional processes. For example, TMS over TPJ has been shown to have effects on visual extinction (Meister et al., 2006), exogenous cueing (Chica et al., 2011), and re-orienting in an attentional capture paradigm (Chang et al., 2013).

cTBS to the pMFG had overall similar effects in the given experiment (line bisection, visual discrimination; bells test). These effects were, however, less systematically corroborated in this experiment than the effects after cTBS to TPJ. Given the limited number of subjects, and the fact, that cTBS does not necessarily induce big behavioural effects in healthy subjects make it difficult to interpret as to why qualitatively similar effects after right pMFG (line bisection, visual discrimination; bells test) were less systematically corroborated in this experiment compared to effects observed after right TPJ cTBS. Lack of statistical power might be one explanation.

Nevertheless, the directional hemi-field effects of cTBS to the right pMFG are important to note, since TMS studies so far rather suggested a symmetric functional role of the right frontal cortex including the pMFG and FEF (Grosbras and Paus, 2002&2003; Chanes et al., 2012; Duecker et al., 2013). Thus, there might be instances where the right pMFG plays a role in spatially orienting visual attention (to the left hemi-field).

Based on data in a stroke population with neglect (Verdon et al., 2010) it had been hypothesized that the right TPJ area could be associated with visual-perceptive processes, and the pMFG with visual exploration. Overall, there was, however, no double dissociation between effects of cTBS to the right TPJ or pMFG in this data set. Thus, while the study results suggest a role of (both of) these areas for some of the spatial visual attentional processes investigated, it did not provide positive evidence for a different role of these areas as had been hypothesized. The data would rather be in line with the assumption that both areas are part of a functional network subserving spatial visual attention.

There had been some indication in this experimental data suggesting a specialised information processing of (left-sided) allocentric spatial features. For one, reaction times for stimuli with gaps on the left side improved during sessions (practice effects), and errors for these stimuli increased after cTBS (of the right MFG). Thus, as observed clinically, (left-sided) allocentric features might be addressed by specialised attentional information processing by the brain in network areas for visual spatial attention.

The approach of applying cTBS to the ventral attention network as performed in this study is novel. The model cited above (Corbetta and Shulman, 2011) assumes that the ventral network is lateralized to the right hemisphere, is not directly involved in spatial visual attention processes, and does not seem to show transcallosal inhibitory dynamics between homologous areas. In this cTBS study, however, it was shown for the right TPJ and less consistently for the right pMFG that these areas of the ventral attention network could well be involved in spatially orienting visual attention (to the left hemi-field). In addition, it was demonstrated that inhibitory cTBS to the right and then left TPJ can have reverse effects on spatial visual attention processes.

Given the fact that brain damage in patients with deficits in spatial visual attention (i.e. neglect) frequently affects these brain areas of the ventral attention network (Verdon et al., 2010), the findings might have therapeutic implications. Noteworthy, the presented data had been obtained with healthy subjects and thus cannot be applied to brain-damaged patients. Nevertheless, there might be parallel lines of reasoning that could be entertained.

According to Kinsbourne’s ‘opponent processor model’, each hemisphere causes a natural attention bias to the contralateral hemifield (Kinsbourne, 1977). Under normal conditions, the two hemispheres are kept in balance due to inter-hemispheric inhibition. In spatial neglect patients, damage to either hemisphere leaves the contralesional intact hemisphere unopposed. As a result of this reduced inhibition, the contralesional hemisphere becomes overactivated and causes an ipsilesional attention bias.

Regarding any influence on a dysfunctional right brain network (e.g. in this experiment after cTBS to the right TPJ or right pMFG) by an (additional) inhibitory stimulation of homologue areas in the left brain network, the presented data suggests that cTBS to the left TPJ could alter the degree of spatial attention bias, with a net effect to change a bias towards the left for a visual-perceptive task (line bisection, visual discrimination).

Thus, the data raise the question whether inhibitory non-invasive brain stimulation (cTBS) to the left TPJ could be a therapeutic candidate for reversing a rightward bias of spatial attention when the right TPJ (as a posterior part of the right ventral attentional system) is dysfunctional as in right brain damaged patients with visual neglect. In a similar vein, one could ask whether inhibitory stimulation (cTBS) to the left pMFG could be a therapeutic candidate for reversing a rightward bias of spatial attention when the right MFG (as a frontal part of the right ventral attentional system) is dysfunctional.

References

Appelros

, Karlsson

G.M.

, Seiger

, & Nydevik

(2002). Neglect and anosognosia after first-ever stroke: Incidence and relationship to disability. Journal of Rehabilitative Medicine, 34, 215–220.

Bowen

, Hazelton

, Pollock

, & Lincoln

N.B.

(2013). Cognitive rehabilitation for spatial neglect following stroke. Cochrane Database Systematic Reviews, 7, CD003586–.

Buxbaum

L.J.

, Ferraro

M.K.

, Veramonti

, Farne

, Whyte

, Ládavas

, et al. (2004). Hemispatial neglect: Subtypes, neuroanatomy, and disability. Neurology, 62, 749–756.

Cazzoli

, Wurtz

, Müri

R.M.

, Hess

C.W.

, & Nyffeler

(2009). Interhemispheric balance of overt attention: A theta burst stimulation study. European Journal of Neuroscience, 29, 1271–1276.

Cazzoli

, Müri

R.M.

, Schumacher

, et al. (2012). Theta burst stimulation reduces disability during the activities of daily living in spatial neglect. Brain, 135, 3426–3439.

Chanes

, Chica

A.B.

, Quentin

, & Valero-Cabré

(2012). Manipulation of pretarget activity on the right frontal eyefield enhances conscious visual perception in humans. PloS One, 7, e36232.

Chang

C.F.

, Hsu

T.Y.

, Tseng

, Liang

W.K.

, Tzeng

O.J.

, Hung

D.L.

, & Juan

C.H.

(2013). Right temporoparietal junction and attentional reorienting. Human Brain Mapping, 34, 869–877.

Chica

A.B.

, Bartolomeo

, & Valero-Cabré

(2011). Dorsal and ventral parietal contributions to spatial orienting in the human brain. Journal of Neuroscience, 31, 8143–8149.

Corbetta

, Kincade

M.J.

, Lewis

, Snyder

A.Z.

, & Sapir

(2005). Neural basis and recovery of spatial attention deficits in spatial neglect. Nature Neuroscience, 8, 1603–1610.

10.

Corbetta

, & Shulman

G.L.

(2011). Spatial neglect and attention networks. Annual Review of Neuroscience, 34, 569–599.

11.

Duecker

, Formisano

, & Sack

A.T.

(2013). Hemispheric differences in the voluntary control of spatial attention: Direct evidence for a right-hemispheric dominance within frontal cortex. Journal of Cognitive Neuroscience, 25, 1332–1342.

12.

Faul

, Erdfelder

, Buchner

, & Lang

A.G.

(2009). Statistical power analyses using G*Power 3.1: Tests for correlation and regression analyses. Methods in Behavioral Research, 41. 1149–1160.

13.

Fierro

, Brighina

, Oliveri

, Piazza

, La Bua

, Buffa

, & Bisiach

(2000). Contralateral neglect induced by right posterior parietal rTMS in healthy subjects. Neuroreport, 11, 1519–1521.

14.

Gauthier

, Dehaut

, & Joanette

(1989). The bells test: A quantitative and qualitative test for visual neglect. International Journal of Clinical Neuropsycholy, 11, 49–53.

15.

Grosbras

M.-H.

, & Paus

(2002). Transcranial magnetic stimulation of the human frontal eyefield: Effects on visual perception and attention. Journal of Cognitive Neuroscience, 14, 1109–1120.

16.

Grosbras

M.H.

, & Paus

(2003). Transcranial magnetic stimulation of the human frontal eyefield facilitates visual awareness. European Journal of Neuroscience, 18, 3121–3126.

17.

Huang

Y.-Z.

, Edwards

M.J.

, Rounis

, Bhatia

K.P.

, & Rothwell

J.C.

(2005). Theta burst stimulation of the human motor cortex. Neuron, 45, 201–206.

18.

Jewell

, & McCourt

M.E.

(2000). Pseudoneglect: A review and meta-analysis of performance factors in line bisection tasks. Neuropsychologia, 38, 93–110.

19.

Kim

B.R.

, Chun

M.H.

, Kim

D.Y.

, et al. (2013). Effect of high- and low-frequency repetitive transcranial magnetic stimulation on visuospatial neglect in patients with acute stroke: A double- blind, sham-controlled trial. Archives of PhysicalMedicine and Rehabilitation, 94, 803–807.

20.

Kinsbourne

(1977). Hemi-neglect and hemisphere rivalry. Advances in Neurology, 18, 41–49.

21.

Loetscher

, Nicholls

M.E.R.

, Brodtmann

, Thomas

N.A.

, & Brugger

(2012). Disentangling input and output-related components of spatial neglect.,, article. Frontiers in Human Neuroscience, 6, article 176.

22.

Marsh

E.B.

, & Hillis

A.E.

(2008). Dissociation between egocentric and allocentric visuospatial and tactile neglect in acute stroke. Cortex, 44, 1215–1220.

23.

Meister

, Wienemann

, Buelte

, Grünewald

, Sparing

, Dambeck

, & Boroojerdi

(2006). Hemiextinction induced by transcranial magnetic stimulation over the right temporo-parietal junction. Neuroscience, 142, 119–123.

24.

Nijboer

T.C.W.

, Kollen

, & Kwakkel

(2013). Time course of visuospatial neglect early after stroke: A longitudinal cohort study. Cortex, 49, 2021–2027.

25.

Nijboer

T.C.W.

, Kollen

, & Kwakkel

(2009). The impact of recovery of visuo-spatial neglect on motor recovery of the upper paretic limb after stroke. PLoS ONE, 9(6), e100584.

26.

Nyffeler

, Cazzoli

, Hess

C.W.

, & Müri

R.M.

(2009). One session of repeated parietal theta burst stimulation trains induces long-lasting improvement of visual neglect. Stroke, 40, 2791–2796.

27.

Ota

, Fujii

, Suzuki

, Fukatsu

, & Yamadori

(2001). Dissociation of bodycentered and stimulus centered representations in unilateral neglect. Neurology, 57, 2064–2069.

28.

Schenkenberg

, Bradford

D.C.

, & Ajax

E.T.

(1980). Line bisection and unilateral visual neglect in patients with neurologic impairment. Neurology, 30, 509–517.

29.

Shirani , et al. (2009). Severity of Hypoperfusion in Distinct Brain Regions Predicts Severity of Hemispatial Neglect in Different Reference Frames. Stroke, 40, 3563–3566.

30.

Verdon

, Schwartz

, Lovblad

K.O.

, Hauert

C.A.

, & Vuilleumier

(2010). Neuroanatomy of hemispatial neglect and its functional components: A study using voxel-based lesion-symptom mapping. Brain, 133, 880–894.