**7. Neuroimaging analysis**

Several brain structures have been associated with PB. In this context, Pedersen et al.(Pedersen et al., 1996) and Santos-Pontelli et al. (Santos-Pontelli et al., 2011) have indicated a wide range of findings from no visible lesion to massive hemispheric lesions on neuroimaging scans in a large sample of PB patients. In these studies, radiologists and neurologists analyzed computed tomography or magnetic resonance imaging in order to determine the type and location of the encephalic lesions. Pedersen at al. determined the size of the stroke by the largest diameter of the lesion (Pedersen et al., 1996).

New Insights for a Better Understanding

needed to confirm this observation.

neuroimaging scans of stroke and non-stroke patients with PB.

**8. Clinical implications of neuroimaging findings** 

interpret the results of the several PB studies.

the 'lesion method' has some noteworthy limitations.

(Ticini et al., 2009).

of the Pusher Behavior: From Clinical to Neuroimaging Features 251

perfusion-weighted imaging (PWI), diffusion-weighted imaging (DWI) and T2-weighted fluid-attenuated inversion-recovery (FLAIR) imaging,. Moreover, they found no damage or malperfusion of the thalamus in patients with PB caused by extra-thalamic lesions. While DWI and FLAIR imaging reveal information about irreversibly damaged neural tissue, PWI allows the identification of structurally intact but not enough to function normally. These interesting findings indicate that the thalamic as well as the extra-thalamic brain structures previously related to the PB contribute to the network controlling upright body posture

Most recently, the relationship between neuroimaging data of stroke and non-stroke PB patients and the severity and prognosis of PB was analyzed (Santos-Pontelli et al., 2011). In order to measure the hemorrhage stroke volume (HSV) in patients with hemorrhagic stroke it was used the ABC/2 method (Zazulia et al., 1999) on CT scans of the acute stroke stage. A positive correlation of the National Institute of Health Stroke Scale (NIHSS) score with HSV in hemorrhagic stroke PB patients was found. In spite of this fact, neither the NIHSS score nor HSV were related with the severity or recovery time of PB. Conversely, previous studies showed that the hemorrhagic volume is highly associated with functional and neurologic deficits (NINDS ICH Workshop Participants, 2005). These data and the fact that the NIHSS score is a good neurologic outcome predictor (Wilde et al, 2010; Wityk et al., 1994; Aslanyan et al., 2004) indicate that the PB evolution and severity may be independent from other neurologic deficits such as those measured by the NIHSS. However, more research is

The fact that all the pusher patients described in literature had an acute event raises the question that the velocity of lesion´s onset may be essential for PB occurrence. In fact, PB also has been reported in patients with other acute brain lesions other than stroke, but not in patients with chronic neurodegenerative disorders (Santos-Pontelli et al., 2004). These observations may indicate that the related alteration of postural control observed in PB may be a consequence of any acute encephalic lesion that lead to a dysfunction in the neural network which processes the input for vertical perception. Figure 3 and 4 show examples of

The analysis of the clinical implications of neuroimaging findings requires an important discussion about some limitations of the neuroimaging methods in order to critically

The localization of human brain functions by studying the correlation between a behavioral disorder and the region of brain lesion has an historical and huge contribution to the understanding of brain function. Nevertheless, as well as all the neuroimaging techniques,

Roden and Karnath pointed out that the lesion method usually assumes that after a focal lesion, the intact regions of the brain continue to function in the same manner as before the lesion (Roden & Karnath, 2004). However, with tasks controlled by spread and changeable circuits, the brain start to adapt rapidly following the lesion. This rearrangement is helpful for recovery, but makes it difficult to infer the original function of the healthy brain. Also, the design of the brain, its blood supply and the surrounding skull mean that some areas of the brain are injured more often than others what implicate that the locations of brain damage are not randomly distributed in the brain. Roden and Karnath highlighted that this

Nevertheless, the location of lesion more consistently described as related to PB occurrence is the posterior thalamus (Karnath et al., 2000a; Karnath et al., 2005). Besides the usual consideration as a relay structure of vestibular pathway (Deecke et al., 1974; Buttner &Henn, 1976), the posterior thalamus is also assumed to be essentially involved in the control of upright body posture. For the lesion analysis, Karnath et al. (Karnath et al., 2000a) compared patients with PB to patients without PB but comparable demographic and clinical data. Using the Talaraich space, the central area of overlap was defined as those voxels in the MRI template that were lesioned in at least 53% or more of their series (number of PB patients=15). The center of lesion overlap was located in the ventral posterior and lateral posterior of the posterolateral thalamus.

Among 40 patients with thalamic strokes (14 pusher patients and 26 control patients), Karnath et al. (Karnath et al., 2005) found that pusher patients had lesions that typically were caused by thalamic hemorrhage. This observation seems to resemble the fact that thalamic hemorrhages predominantly affect the posterolateral part of the thalamus (Hungerbuhler et al., 1984; Kawahara et al., 1986; Kumral et al., 1995; Chung et al., 1996) and that infarctions are less frequent in the posterior thalamus vs. the anterior and paramedian thalamus (Bogousslavsky et al., 1988; Van der Werf et al., 2000). Nevertheless, their control patients presented more ischemic than hemorrhagic thalamic strokes. Using two standard protocols, the authors carried out MRI or spiral CT imaging that were fit the canonical AC-PC orientation of the MRI scans (Karnath et al., 2005). The boundary of the lesion was delineated directly on the individual MRI for every single transversal slice using MRIcro software (Roden & Brett, 2000). Both the scan and lesion shape were then transferred into stereotaxic space using the spatial normalization algorithm provided by SPM2. The MRIcro software was also used to map the lesion from transversal slices of the T1-template MRI from Montreal Neurologic Institute (MNI) space. The authors used the Talairach Zcoordinates in Talairach space by using the identical or the closest matching transversal slices of each individual. Lesion location in the thalamic stroke patients with and without PB was compared using the subtraction technique (Karnath et al., 2005). The percentage of overlapping lesions of the PB patients after subtraction of controls ranged 20%.

The PB is also observed in patients with brain lesions that spare the thalamus as postcentral gyrus (Johannsen et al., 2006b), internal capsule (Pedersen et al., 1996; Saj et al., 2005), temporal lobe (Pedersen et al., 1996; Johannsen et al., 2006b), supplementary motor area (Reding et al., 1997), superior parietal lobule (Reding et al., 1997), inferior parietal lobule (Johannsen et al., 2006b), globus pallidus (Reding et al., 1997), striatum (Saj et al., 2005), centrum semi-ovalum (Saj et al., 2005), insula (Reding et al., 1997; Johannsen et al., 2006b), isolated cerebellum (Paci &Nannetti, 2005) and isolated anterior cerebral artery territory (Karnath et al., 2008).

By analyzing neuroimaging scans of patients with and without PB with the same methodology of Karnath et al. (Karnath et al., 2005), Johannsen et al. found very small regions for pusher patients when subtracted from matched controls (Johannsen et al., 2006b). In both hemispheres, the lesion of the pusher patients centered at the insular cortex and the postcentral gyrus. However, these areas were identified with the subtraction technique where the percentage of difference between the pusher and control patients neuroimaging scans was not exclusively 100% (ranged from 81 to 100%). Although both were meticulous studies (Karnath et al., 2005; Johannsen et al., 2006b), this analysis does not exclude the same lesion location in control patients.

Recently, Ticini et al. (Ticini et al., 2009), found that the posterior thalamus itself is integral to the occurrence of PB rather than additional malperfusion in distant cortical areas by using

Nevertheless, the location of lesion more consistently described as related to PB occurrence is the posterior thalamus (Karnath et al., 2000a; Karnath et al., 2005). Besides the usual consideration as a relay structure of vestibular pathway (Deecke et al., 1974; Buttner &Henn, 1976), the posterior thalamus is also assumed to be essentially involved in the control of upright body posture. For the lesion analysis, Karnath et al. (Karnath et al., 2000a) compared patients with PB to patients without PB but comparable demographic and clinical data. Using the Talaraich space, the central area of overlap was defined as those voxels in the MRI template that were lesioned in at least 53% or more of their series (number of PB patients=15). The center of lesion overlap was located in the ventral posterior and lateral

Among 40 patients with thalamic strokes (14 pusher patients and 26 control patients), Karnath et al. (Karnath et al., 2005) found that pusher patients had lesions that typically were caused by thalamic hemorrhage. This observation seems to resemble the fact that thalamic hemorrhages predominantly affect the posterolateral part of the thalamus (Hungerbuhler et al., 1984; Kawahara et al., 1986; Kumral et al., 1995; Chung et al., 1996) and that infarctions are less frequent in the posterior thalamus vs. the anterior and paramedian thalamus (Bogousslavsky et al., 1988; Van der Werf et al., 2000). Nevertheless, their control patients presented more ischemic than hemorrhagic thalamic strokes. Using two standard protocols, the authors carried out MRI or spiral CT imaging that were fit the canonical AC-PC orientation of the MRI scans (Karnath et al., 2005). The boundary of the lesion was delineated directly on the individual MRI for every single transversal slice using MRIcro software (Roden & Brett, 2000). Both the scan and lesion shape were then transferred into stereotaxic space using the spatial normalization algorithm provided by SPM2. The MRIcro software was also used to map the lesion from transversal slices of the T1-template MRI from Montreal Neurologic Institute (MNI) space. The authors used the Talairach Zcoordinates in Talairach space by using the identical or the closest matching transversal slices of each individual. Lesion location in the thalamic stroke patients with and without PB was compared using the subtraction technique (Karnath et al., 2005). The percentage of

overlapping lesions of the PB patients after subtraction of controls ranged 20%.

The PB is also observed in patients with brain lesions that spare the thalamus as postcentral gyrus (Johannsen et al., 2006b), internal capsule (Pedersen et al., 1996; Saj et al., 2005), temporal lobe (Pedersen et al., 1996; Johannsen et al., 2006b), supplementary motor area (Reding et al., 1997), superior parietal lobule (Reding et al., 1997), inferior parietal lobule (Johannsen et al., 2006b), globus pallidus (Reding et al., 1997), striatum (Saj et al., 2005), centrum semi-ovalum (Saj et al., 2005), insula (Reding et al., 1997; Johannsen et al., 2006b), isolated cerebellum (Paci &Nannetti, 2005) and isolated anterior cerebral artery territory

By analyzing neuroimaging scans of patients with and without PB with the same methodology of Karnath et al. (Karnath et al., 2005), Johannsen et al. found very small regions for pusher patients when subtracted from matched controls (Johannsen et al., 2006b). In both hemispheres, the lesion of the pusher patients centered at the insular cortex and the postcentral gyrus. However, these areas were identified with the subtraction technique where the percentage of difference between the pusher and control patients neuroimaging scans was not exclusively 100% (ranged from 81 to 100%). Although both were meticulous studies (Karnath et al., 2005; Johannsen et al., 2006b), this analysis does not

Recently, Ticini et al. (Ticini et al., 2009), found that the posterior thalamus itself is integral to the occurrence of PB rather than additional malperfusion in distant cortical areas by using

posterior of the posterolateral thalamus.

(Karnath et al., 2008).

exclude the same lesion location in control patients.

perfusion-weighted imaging (PWI), diffusion-weighted imaging (DWI) and T2-weighted fluid-attenuated inversion-recovery (FLAIR) imaging,. Moreover, they found no damage or malperfusion of the thalamus in patients with PB caused by extra-thalamic lesions. While DWI and FLAIR imaging reveal information about irreversibly damaged neural tissue, PWI allows the identification of structurally intact but not enough to function normally. These interesting findings indicate that the thalamic as well as the extra-thalamic brain structures previously related to the PB contribute to the network controlling upright body posture (Ticini et al., 2009).

Most recently, the relationship between neuroimaging data of stroke and non-stroke PB patients and the severity and prognosis of PB was analyzed (Santos-Pontelli et al., 2011). In order to measure the hemorrhage stroke volume (HSV) in patients with hemorrhagic stroke it was used the ABC/2 method (Zazulia et al., 1999) on CT scans of the acute stroke stage. A positive correlation of the National Institute of Health Stroke Scale (NIHSS) score with HSV in hemorrhagic stroke PB patients was found. In spite of this fact, neither the NIHSS score nor HSV were related with the severity or recovery time of PB. Conversely, previous studies showed that the hemorrhagic volume is highly associated with functional and neurologic deficits (NINDS ICH Workshop Participants, 2005). These data and the fact that the NIHSS score is a good neurologic outcome predictor (Wilde et al, 2010; Wityk et al., 1994; Aslanyan et al., 2004) indicate that the PB evolution and severity may be independent from other neurologic deficits such as those measured by the NIHSS. However, more research is needed to confirm this observation.

The fact that all the pusher patients described in literature had an acute event raises the question that the velocity of lesion´s onset may be essential for PB occurrence. In fact, PB also has been reported in patients with other acute brain lesions other than stroke, but not in patients with chronic neurodegenerative disorders (Santos-Pontelli et al., 2004). These observations may indicate that the related alteration of postural control observed in PB may be a consequence of any acute encephalic lesion that lead to a dysfunction in the neural network which processes the input for vertical perception. Figure 3 and 4 show examples of neuroimaging scans of stroke and non-stroke patients with PB.
