OUP user menu

Left frontal transcranial magnetic stimulation reduces contralesional extinction in patients with unilateral right brain damage

M. Oliveri, P. M. Rossini, R. Traversa, P. Cicinelli, M. M. Filippi, P. Pasqualetti, F. Tomaiuolo, C. Caltagirone
DOI: http://dx.doi.org/10.1093/brain/122.9.1731 1731-1739 First published online: 1 September 1999


It has been demonstrated previously that transcranial magnetic stimulation (TMS) of the sensorimotor cortex can induce transient suppression of the perception of cutaneous near-threshold stimuli from fingers of the contralateral hand in normal individuals. One explanation accounting for deficits in the exploration of contralateral space following a unilateral hemispheric lesion refers to a loss of the normal interhemispheric balance, with a resultant hyperactivation of the unaffected hemisphere due to the release of reciprocal inhibition by the affected one. In order to verify this hypothesis, we investigated the effects of a TMS-induced transient dysfunction of the normal hemisphere upon contralateral tactile extinctions in two groups: (i) 14 right brain-damaged patients and (ii) 14 left brain-damaged control patients. Single-pulse TMS was delivered to frontal and parietal scalp sites of the unaffected hemisphere after an interval of 40 ms from an electrical unimanual or bimanual digit stimulation. In right brain-damaged patients, left frontal TMS significantly reduced the rate of contralateral extinctions compared with controls. After left parietal TMS, the number of extinctions was comparable to the baseline. This pattern of increased sensitivity to cutaneous stimulation ipsilateral to TMS was not observed in left brain-damaged control patients. In this group, right hemisphere TMS did not significantly alter the recognition of bimanual stimuli delivered to the space contralateral to the lesion. The suggestion is made that extinctions produced by right brain damage may be dependent on a breakdown in the balance of hemispheric rivalry in directing spatial attention to contralateral hemispace, so that the unaffected hemisphere generates an unopposed orienting response to the side of the lesion. The mechanisms whereby the left frontal TMS transiently ameliorates these deficits may involve stimulus-induced removal of a left frontal–right parietal transcallosal inhibitory flow, although interactions at subcortical levels cannot be excluded.

  • transcranial magnetic stimulation
  • neglect
  • hemispheric asymmetry
  • right brain damage
  • LBD = left brain-damaged
  • RBD = right brain-damaged
  • TMS = transcranial magnetic stimulation


Directed spatial attention is subserved by a distributed anatomo-functional neural network relaying in the parietal and frontal cortices of both hemispheres (Mesulam, 1990). Damage to specialized circuits of this system within one cerebral hemisphere (more often the right hemisphere) is known to give rise to clinical symptoms such as extinction and neglect. Spatial hemineglect is a failure to recognize and explore stimuli in the hemispace opposite the lesion, including the patient's contralesional body parts (Heilman et al., 1987; Bisiach and Vallar, 1988; Vallar, 1998). The term extinction refers to the phenomenon whereby patients with unilateral brain damage fail to report a stimulus delivered to the side contralateral to the lesion when a symmetrical ipsilateral stimulus is given simultaneously (Vallar, 1998; Gainotti et al., 1989). Although distinct phenomena, contralateral extinction and neglect are often expressions of the same underlying deficit; they are most frequent and severe following lesions to the right cerebral hemisphere, especially of the posterior parietal lobe, which is involved in the representation of personal and extrapersonal space (Heilman et al., 1987; Bisiach and Vallar, 1988; Vallar, 1998). By extension, a number of studies suggest that all the contralateral somatosensory deficits associated with lesions of the right hemisphere (even in the absence of neglect or extinction) are due to the impairment of both primary sensory and attentional components (Vallar et al., 1993b).

Current interpretations of these deficits include a defective conscious representation of contralesional hemispace (Bisiach et al., 1981; Heilman et al., 1987), an ipsilesional deviation of the egocentric frame of spatial reference (such as the perceived mid-sagittal plane) (Jeannerod and Biguer, 1987) and a polarized disuniformity of the medium for the representation of spatial relationships, implying a left–right anisometry of space representation (Bisiach et al., 1994, 1996). All these current models assign a major role to the right hemisphere in attending to and representing both sides of space, while the left hemisphere would be concerned mainly with the contralateral right side (Heilman et al., 1987, 1993; Weintraub and Mesulam, 1987).

On the other hand, another model of the neurological basis of spatial attention postulates the existence in normal individuals of a dynamic balance—with left-hemisphere prevalence—between the two hemispheres in orienting attention towards the contralateral hemispace (Kinsbourne, 1977, 1993, 1994). According to this account, neglect or extinction results from a unilateral lesion because of a breakdown in the balance of hemispheric rivalry, so that the unaffected hemisphere (usually the left one) becomes hyperactivated following the release of reciprocal inhibition by the affected hemisphere. This can have two consequences that are not mutually exclusive: from one side the non-lesioned hemisphere could generate an unopposed orienting response towards the side of the lesion; from the other side, the hyperactivation of attentional neurons of the unaffected hemisphere could lead to enhanced transcallosal inhibition of homologous areas of the affected hemisphere. Both phenomena would lead to a deficit in exploration of contralateral space. This model stresses the need to reconsider the productive (as opposed to defective) manifestations of neglect, and emphasizes the role of the unaffected hemisphere in producing the attentional imbalance. One implication of this theory is that some deficits in the exploration of contralateral space could be reduced by lesions or manoeuvres known to interfere with the function of the unaffected hemisphere.

We addressed this problem by using transcranial magnetic stimulation (TMS) as a technique able to produce focal, transient and fully reversible disruption of cortical network function during the performance of cognitive tasks (Day et al., 1989; Pascual-Leone et al., 1991; Seyal et al., 1992, 1995; Amassian et al., 1993a, b; Gerloff et al., 1997; Walsh and Cowey, 1998).

In the present study we tested for the disrupting effects of TMS on the unaffected hemisphere in right brain-damaged patients compared with left brain-damaged control patients, in a task of discrimination of unimanual and bimanual stimuli delivered to the space ipsilateral or contralateral to the hemispheric lesion.

Patients and methods

Subjects (Table 1)

View this table:
Table 1

Clinical features of the patients

CaseAge (years)SexTime of onset (months)NeglectExtinctionLesion sideLesion site
F = frontal; P = parietal; IC = internal capsule; SC =sub-cortical; T = temporal; O = occipital.

Fourteen right brain-damaged (RBD) and 14 left brain-damaged (LBD) patients with radiological (CT or MRI) and clinical evidence of unilateral brain lesions entered the study. Twenty-seven patients had suffered an ischaemic or haemorrhagic stroke and one patient had a cerebral tumour (case 1). None of the patients had a history or evidence of previous cerebrovascular disease, dementia or psychiatric disorders. They were all examined during a rehabilitation period, within 1–4 months from the onset of symptomatology. The two groups of patients did not differ in mean age, sex or duration of illness.

All subjects were right-handed, according to the Edinburgh inventory (Oldfield, 1971). They gave informed consent for participation in the study, which was approved by the IRCCS S. Lucia ethical committee.

All patients were given a standard clinical neurological and neuropsychological examination to assess motor, somatosensory deficits and the presence of neglect, language disturbances or cognitive impairment.

Six out of 14 RBD patients had evidence of visuospatial hemineglect (N+); contralateral extinctions of tactile stimuli were present in five RBD patients, three of whom also had hemineglect. The remaining six RBD patients showed contralateral somatosensory deficits. Two LBD patients showed contralateral extinctions; none of them had neglect or severe aphasia (see below for the method of assessment of tactile extinction).

Tactile stimulation

During the experiment, the subjects sat comfortably on an armchair with their hands supinated.

Thirty unimanual electrical stimuli (15 to the right-hand digits and 15 to the left-hand digits), randomly intermingled with 15 bimanual symmetrical electrical stimuli (to homologous digits), were delivered with pairs of surface electrodes applied around the first, third and fifth fingers of each hand (cathode on the second phalanx, anode on the third phalanx). Square-wave pulses, 0.3 ms in duration, were discharged by a dedicated electronic device. Stimulus intensity was initially subthreshold; the intensity of stimulation was then gradually increased until it was set ~10% above the subjective threshold for perception. Finger stimuli were separated by a fixed interval in a randomized sequence.

Patients who perceived nearly all contralateral single stimuli but failed to perceive the contralateral stimulus during bimanual stimulation in >50% of the trials were considered as showing extinction. Patients who failed to report 30–100% of single stimuli delivered to the contralateral hand were considered to show somatosensory deficits.

Transcranial magnetic stimulation

TMS was performed with a Novametrix MagStim 200 magnetic stimulator (MagStim Co. Ltd, Withland, UK) using a figure-of-eight coil (double 70-mm coil, MagStim). The coil was placed tangential to the skull, with the handle pointing backwards parallel to the midline. This induced a current flowing in a posterior–anterior direction in the underlying brain areas.

The excitability threshold was defined, in accordance with international standards, as the lowest stimulus intensity required to elicit motor evoked potentials of >50 V peak-to-peak amplitude in the contralateral abductor pollicis brevis muscle in at least five of 10 trials with the coil centred over the optimal scalp position (Rossini et al., 1994).

Then, after an interval of 40 ms from the presentation of stimuli to the finger(s), single-pulse TMS was applied at an intensity 10% above the subject's unaffected hemisphere motor threshold to scalp positions centred on the left/right frontal and parietal sites (corresponding to the F3, F4, P3 and P4 labels of the 10–20 EEG system) of the unaffected hemisphere. This interval was selected as the most active one in inducing sensory disruption in a group of healthy controls.


Following the presentation of individual stimulus pairs (tactile + transcranial stimulus), the subject was asked to indicate whether he or she perceived the tactile stimulus or stimuli and to localize the stimulus or stimuli.

The subjects performed the task in four blocks of 45 stimulus pairs for each scalp site. The order of the stimulated scalp positions was randomized across subjects. During the task the eye/head gaze was directed straight ahead.

Control trials

Baseline trials (finger stimuli not followed by the interfering TMS), randomly intermingled with test conditions, were gathered with the magnetic coil still held over the scalp, to control for non-specific effects of the experimental procedure (i.e. learning, spatial attention). Trials in which TMS was applied without the `target' finger stimulation were used as a measure of the subject's state of alertness.

In order to test for any possible non-specific effects on the perception of cutaneous stimuli of the magnetic pulse-associated sound, eye-blinking and head-muscle contraction due to spreading of the stimulus to the trigeminal and/or facial nerves, a focal TMS was also applied to scalp positions centred on the prefrontal area of the unaffected hemisphere (Fp1, Fp2). In all subjects, stimulation over these control positions was randomized with the parietal and frontal stimuli.


Error rates, as referred both to incorrect (stimulus located to a different digit or the contralateral hand) and omitted responses (stimulus suppression) of stimuli contralateral or ipsilateral to the affected hemisphere, were analysed in the different experimental conditions.

Statistical analysis was performed with a three-way analysis of variance of mean response errors, evaluating the effects of group (RBD versus LBD patients) as the between-subject factor, and of type of error (four levels: contralateral errors during unimanual stimulation; contralateral errors during bimanual stimulation; ipsilateral errors during unimanual stimulation; ipsilateral errors during bimanual stimulation) and condition (four levels: baseline, frontal TMS, parietal TMS, prefrontal TMS) as within-subject factors. Greenhouse–Geisser correction to degrees of freedom was applied when appropriate. Post hoc comparisons (with Tukey's correction) were used to test differences between levels of the different factors. The threshold of significance was set at P < 0.05.

MRI scan (Fig. 1)

Fig. 1

MRI scan of patient 6. The left side of the figure shows the three-dimensional reconstruction (upper part) and the MRI sagittal scan (lower part) of the unaffected (left) hemisphere. On the sagittal scan, two lines were drawn tangential to the skull at the points of TMS (frontal and parietal; white circle with black asterisk), dropping a perpendicular into the grey matter (black circle with white asterisk). For better visualization of the sites of stimulation we also reproduce them on the three-dimensional surface of the brain. The frontal site was in the frontal lobe over the gyrus frontalis secundus. The parietal site was in the parietal lobe over the gyrus centralis posterior. The right side of the figure shows the reconstruction of the lesion (wireframe) on a three-dimensional view of the right hemisphere (upper part) and the corresponding MRI sagittal scan (see text for details). f2 = inferior frontal sulcus; fS = Sylvian fissure; R = central sulcus; Spo = sulcus postcentralis; t1 = sulcus temporalis.

In a single patient (patient 6), who had right hemisphere damage, the frontal and parietal sites of the unaffected hemisphere were marked on the skull using capsules containing soya oil. A set of T1-weighted images was produced with a Siemens 1.5 T Vision Magnetom MR system (Erlangen, Germany; MPRAGE sequence, 1 mm isotropic voxels). The MRI scans were reconstructed following the sagittal line passing through the centre of both capsules and perpendicular to the skull; thereafter, a line parallel to the capsule and tangential to the surface of the skull and a line perpendicular to this line, originating in the centre of the capsule, were drawn (see MRI sagittal scan of the left hemisphere; Fig. 1). These lines indicate, respectively, the coil orientation and the centre of the area where the induced field was at its maximum.

The external landmarks represented by the capsules and the underlying brain cortex were identified with DISPLAY (Brain Imaging Center, Montreal Neurological Institute, McGill University, Canada). DISPLAY permits demarcation of a region of interest within the MRI volume and allows three-dimensional visualization. The brain surface was created using a three-dimensional model-based surface deformation algorithm (MacDonald, 1998).

The stimulated scalp sites were localized by the method of Economo and Koskinas (1925), after having identified pertinent sulcal landmarks on the left hemisphere. We localized the frontal site above the inferior frontal sulcus (f2), in the gyrus frontalis secundus (Brodmann areas 9 and 46) and the parietal site in the parietal lobe, among the central sulcus (R) and the sulcus postcentralis [Spo; Brodmann areas 1 and 2] (see left side of Fig. 1).

The right side of Fig. 1 shows the three-dimensional reconstruction of the lesion (wireframe) and its MRI sagittal view. The lesion extended from the sylvian fissure (fS) to the sulcus temporalis superior (t1) dorsal and ventral bank, including the gyrus temporalis primus and part of the gyrus temporalis secundus up to the temporal pole. The lesion also involved the inferior portion of the pars opercularis and pars tiangularis, the operculum Rolando, the operculum parietale and the lobus supramarginalis.


Motor thresholds of the non-affected hemisphere did not differ between RBD and LBD patients (52.33 ± 7.31 versus 48 ± 4.40%; P > 0.05).

During the different experimental conditions, the patients' reports of incorrect responses (stimulus dislocation) were virtually absent (<2% of all trials). Accordingly, only response omissions were considered for statistical analysis in RBD and LBD patients.

Figure 2 shows the mean percentage of contralateral and ipsilateral errors during the different experimental conditions in the two groups of patients. Inspection of Fig. 2 clearly reveals a marked difference in performance between the RBD and LBD patients, substantiated by the significance of the triple interaction group × type of error × condition [F(3.9,92.3) = 2.5; P = 0.050].

Fig. 2

Mean percentage of contralateral and ipsilateral errors according to peripheral discrimination task (uni- and bimanual) and TMS scalp positions in right brain-damaged (open circles) and left brain-damaged (filled squares) patients.

When considering contralateral errors during unimanual discrimination tasks (Fig. 2A), the significance of the interaction condition × group [F(1.8,46.8) = 7.76; P = 0.002] reflected the fact that right frontal and parietal TMS significantly increased the rate of errors (Tukey's P < 0.0001) in LBD patients, whereas in RBD patients the percentage of omissions after TMS of the unaffected hemisphere was similar to that in baseline trials (P > 0.05).

The most interesting result is the different pattern of recognition of contralateral stimuli observed in the two groups during bimanual discrimination tasks [F(1.8,43.9) = 9.01; P = 0.001; Fig. 2B]. In particular, after frontal TMS, in RBD patients there was a marked decrease in detection misses, whereas in the LBD group no significant change in the number of contralateral extinctions was induced by right hemisphere TMS.

In addition, post hoc comparisons showed that in baseline trials during bimanual stimulation the mean percentage of contralateral extinctions was significantly higher in RBD than in LBD patients (83.09 ± 21.93 and 61.31 ± 30.98%, respectively; Tukey's P = 0.04).

At individual analysis, 13 out of 14 (93%) RBD patients showed improvement of the contralateral extinctions after left hemisphere TMS. The positive effects of TMS in these patients were not confined to those with neglect or contralateral extinction: the contralateral somatosensory deficits of five out of six RBD patients without clinical evidence of neglect or extinction were also positively affected by left frontal TMS. The performance on this task of the RBD patients with and without neglect and/or contralateral extinctions is shown in Table 2.

View this table:
Table 2

Mean percentage of contralateral extinctions in right brain–damaged patients during the different experimental conditions

BaselinePrefrontal TMSFrontal TMSParietal TMS
N+: patients with neglect; Ext+ = patients with extinction; SD = patients with contralateral somatosensory deficits.
N+ (n = 6)94.62 ± 4.5890.81 ± 6.6584.44 ± 5.4496.38 ± 4
Ext+ (n = 5)74.44 ± 26.0572.08 ± 24.1650.47 ± 27.4770.83 ± 26.41
SD (n = 6)70.12 ± 24.0772.77 ± 21.3351.74 ± 27.6475.55 ± 25.53

In all 14 LBD patients, the contralesional missed stimuli were unchanged after right hemisphere TMS.

When considering ipsilateral errors (Fig. 2C and D) we found significant main effects for type of error [F(1.3) = 7.23; P = 0.012] and for condition [F(1.9,49.4) = 8.95; P = 0.001]. No significant differences were found between RBD and LBD patients either for unimanual or bimanual discrimination tasks [F(2.1,54.9) = 0.24; P = 0.803 and F(1.8, 45.8) = 0.131; P = 0.853, respectively].

These findings indicate that stimulation of frontal and parietal sites of the unaffected hemisphere significantly altered the recognition of stimuli delivered to the ipsilesional (i.e. contralateral to the TMS site) space, but this interference did not differ significantly between RBD and LBD patients. Moreover, the number of ipsilateral extinctions (reflecting the ipsilesional orienting response) in baseline conditions was similar in all patients (17.62 ± 10.91% in RBD patients with neglect; 23.88 ± 11.38% in RBD patients with extinction; 27.03 ± 20.09% in RBD patients without neglect or extinction; 27.69 ± 19.91% in LBD patients).

Prefrontal TMS did not significantly alter the pattern of recognition of contralateral or of ipsilateral (unimanual or bimanual) stimuli in either group of patients compared with baseline trials.

Finally, no clear-cut differences were found in the site and extent of lesion between patients influenced and not influenced by TMS.


The main finding of the present study was that, in patients with right hemisphere damage, single-pulse TMS to the left frontal cortex, applied at a suitable time after a bimanual discrimination task, reduced the number of extinctions of tactile stimuli delivered to the contralateral hemispace. The phenomenon was not limited to patients with neglect or contralateral extinction but was also evident in those showing only contralateral somatosensory deficits. On the other hand, in patients with left-hemisphere damage, TMS of the unaffected (right) hemisphere did not significantly alter the recognition of stimuli from the contralateral hemispace.

Since the unaffected hemisphere motor thresholds did not differ significantly between the two groups of patients, these results cannot be attributed to different TMS intensities relative to different thresholds of excitability. In addition, the topographic specificity of the phenomenon, evident after frontal but not parietal or prefrontal left hemisphere TMS, seems to rule out any explanation of extinction recovery based on unspecific effects of the TMS-associated sound, blinking or contraction of scalp muscles, shifting attention towards the stimulated side.

Our findings support the idea that perceptual deficits in the exploration of contralateral hemispace after a right-hemisphere lesion occur not only because of the loss of the right hemisphere's ability to attend to the left hemispace, but also because of an imbalance between competing systems for lateral orientation of the two hemispheres (Kinsbourne, 1977, 1993, 1994). One variant of this theoretical hemispheric rivalry emphasizes the notion of mutually inhibitory callosal connections between the two hemispheres. According to this account, when one hemisphere is lesioned, homologous regions of the opposite hemisphere, which normally receive inhibitory projections from the damaged hemisphere, become disinhibited and can therefore hyperorient attention to the ipsilesional side and/or further inhibit the affected hemisphere.

The current investigation provides some support to this theory. In fact, the improvement in contralateral extinctions in RBD patients could be due to transient TMS-induced disruption of the underlying areas of the unaffected hemisphere, resulting in recovery of the interhemispheric imbalance in conscious sensory perception created by the stroke.

On the other hand, no significant differences in the mean percentage of ipsilateral (to the affected hemisphere) errors were found between RBD and LBD patients. This finding seems to argue against the other assumption of Kinsbourne's model (Kinsbourne, 1977, 1993) of abnormally focused attention towards the right (ipsilesional) side as the dominant factor accounting for contralateral space perception deficits, at least when considering the performance of all RBD patients. In effect, a lower (though not significantly lower) number of ipsilateral extinctions was noted in RBD patients with neglect compared with those without neglect or in LBD patients. On the other hand, the limited number of patients included in the various groups does not allow us to make definitive conclusions on this topic.

The observed topographic specificity of the TMS over the unaffected hemisphere in inducing recovery from contralateral extinctions is in accord with previous clinical observations of neglect recovery in RBD patients following a second lesion occurring in the left frontal lobe (Vuilleumier et al., 1996), while a definite worsening followed a second lesion in the left parietal areas (Pierrot-Deseilligny et al., 1986). This clinical evidence and the results of the present study can both be explained by the existence of a dominant callosal inhibitory influence directed from the left frontal lobe to the right parietal areas after being relayed in homologous right frontal regions; this view would be in accord with the proposed theory that, in right-handed patients, the mutual transcallosal inhibition between the attentional neurons of the two hemispheres is asymmetrical, with more prominent inhibition directed from the dominant to the non-dominant hemisphere (Kinsbourne, 1977). Therefore, a disruption of the left frontal cortex processing, whether permanent (stroke) or transient (following single-pulse TMS), would interfere with this frontoparietal inhibition, with the net effect of a right parietal disinhibition and consequent enhanced detection of stimuli from the left hemispace (Fig. 3). On the other hand, a disruption of the left parietal cortex would interfere only with the pathological hyperorientation of the left attentional neurons to the right hemispace, its effects on extinction recovery therefore being comparatively minor or absent.

Fig. 3

Theoretical frame of a possible pattern of left- and right-hemisphere contributions to the overall neural representation of egocentric space. (A) In normal individuals, the mutual inhibitory callosal connections between the two hemispheres are asymmetrical: inhibitory connections from the dominant to the non-dominant hemisphere are stronger, thereby determining a slight hyperorientation of attention towards the right side. (B) After a right hemisphere stroke, the imbalance between competing systems for lateral orientation results in excessive attention towards the right (ipsilesional) hemispace (dashed arrow), due to the unbalanced effect of the left hemisphere. (B) In right brain-damaged patients, left frontal TMS would interfere with a hypothetical left frontal–right parietal inhibition vector, with the net effect of a right parietal disinhibition and consequent partial restoration of left extinctions (black arrow).

A limit intrinsic to the TMS technique is its topographic specificity in identifying the anatomical sites stimulated. In this study we stimulated two scalp sites identified on the basis of the 10–20 EEG system. Co-registration of these sites with anatomical MRI, performed in a representative patient, revealed that the frontal site was above the inferior frontal sulcus (f2), in the gyrus frontalis secundus (Brodmann areas 9 and 46). These regions correspond, at least in part, to those whose damage was associated with the neglect recovery described in the study reported previously (Vuilleumier et al., 1996). It is well known that these regions are some of the major components of the cortical circuitry subserving spatial attention in humans (Mesulam, 1990).

As opposed to what was seen in RBD patients, in LBD patients the right hemisphere TMS was not effective in producing disinhibition of the opposite (damaged) hemisphere and it did not result in improved recognition of contralateral sensory stimuli. This non-symmetrical pattern of TMS effects can be interpreted in terms of hemispheric differences in the processing of somatosensory stimuli. It has been assumed that in LBD patients contralateral somatosensory deficits mainly reflect damage to the primary sensory processing of the stimulus, primarily performed by the contralateral anterior parietal lobe, without left–right asymmetries (Vallar et al., 1993). By contrast, in RBD patients, regardless of the presence of neglect or extinction, contralateral somatosensory deficits would involve both sensory and attentional factors. In fact, the processes related to conscious perception of somatosensory stimuli are organized in a non-symmetrical fashion, with right-hemisphere dominance, at the level of the posterior parietal lobe (Vallar et al., 1990, 1993b). It is therefore conceivable that a manoeuvre such as TMS, which, by disrupting the interhemispheric distribution of attentional information, would affect the more central levels of information processing, was effective only in RBD patients.

It is also worth noting that in RBD patients only contralateral extinctions—primarily produced by a deficit at the attentional level—were influenced by left hemisphere TMS, whereas contralateral unimanual stimuli were not influenced. This finding confirms that the contralateral deficits ameliorated by TMS disruption of the disinhibited healthy hemisphere have an attentional component.

Experimental studies in animals suggest that both cortical and subcortical mechanisms can contribute to the reversal of spatial attentional bias observed in RBD patients. In fact, it is known that cortical lesions also produce an imbalance in the activity of subcortical structures involved in orienting attention, such as the superior colliculus (Sprague, 1966). Parieto-occipital projections to the ipsilateral superior colliculus normally exert a tonic facilitatory drive. After a unilateral cortical lesion, the superior colliculus loses this tonic activation and the opposite (contralesional) superior colliculus becomes hyperactive due to increased activation from its disinhibited hemisphere. If the contralesional superior colliculus is then removed, the hyperorientation, and hence neglect, is ameliorated (Wallace et al., 1989). However, this effect seems to be valid especially for the visual system, and so does not apply significantly to our protocol, in which space exploration deficits are investigated with a tactile modality.

Evidence from primates shows that transcallosal connections between the primary somatosensory areas of distal body parts are sparse compared with the numerous connections between representations of proximal body parts (Rouiller et al., 1994). Moreover, it is not known whether the interhemispheric connections between distal limb representations are predominantly facilitatory or inhibitory. However, in the present study we focused our attention on callosal fibres interconnecting the hand representations of associative parietal and frontal areas; these callosal connections, as well as those connecting the non-primary motor areas in the frontal lobe, are more powerful and widespread than those between the primary hand motor and sensory areas (Rouiller et al., 1994). In addition, there has been a progressive increase in interhemispheric connections during phylogenesis from monkey to humans (Meyer et al., 1998). Thus, TMS of the frontal site might be expected to have a pronounced effect on transcallosal fibres that may account for the present results. Obviously, other routes could be involved in parallel or as an alternative to the transcallosal route. They include (i) direct projections from the ventrolateral thalamic nuclei to frontal lobes bypassing the primary somatosensory cortex, (ii) ipsilateral projections to the spinal cord from the frontal lobes, and (iii) crossed spinal interneuron circuits which could potentially modulate ascending inputs and effects from the motor areas upon sensation from the ipsilateral limb. The present protocol does not allow us to discriminate between the contributions of cortical and subcortical brain regions for these interactions.

In conclusion, the results of our study provide for the first time neurophysiological evidence of asymmetrical mutual inhibition between the attentional neurons of the two hemispheres in right-handed humans, similar to that previously reported for the motor areas (Netz et al., 1995). These findings stress the important role played by the cortical commissures in the distribution of attentional resources between the cerebral hemispheres and suggest that a basic deficit of hemineglect involves hyperactivation of the unaffected hemisphere due to the release of reciprocal inhibition by the affected hemisphere. However, they do not rule out other interpretations of this multifaceted syndrome, which includes multiple components and a number of symptoms and signs resulting from the disordered function of discrete neural networks.


View Abstract