The Neural Basis of Anosognosia for Spatial Neglect After Stroke
Background and Purpose—The present study investigated the lesion anatomy of anosognosia for visuospatial neglect resulting from right hemispheric stroke.
Methods—In 63 patients, self-ratings of performance in paper-and-pencil tests were contrasted with external performance ratings. Lesion analysis was conducted on patient subgroups with different degrees of anosognosia but comparable visuospatial impairment.
Results—Independent of the severity of visuospatial neglect per se, damage to the right angular and superior temporal gyrus was associated with higher levels of anosognosia.
Conclusions—Using a novel assessment of anosognosia for spatial neglect, the present study relates stroke-induced self-awareness deficits to inferior parietal and superior temporal brain damage.
Despite the severe impact of stroke-induced impairments on daily life, some patients remain unaware of their deficits, that is, show anosognosia.1 Previous studies demonstrated that the majority of patients with reduced self-awareness for hemiplegia/hemiparesis have right hemispheric brain damage.2,3 No study has yet focused on the lesion anatomy of self-awareness deficits for visuospatial neglect, although these are even more prevalent than for hemiplegia.4 The present study investigated anosognosia for spatial neglect and its neuroanatomy in patients with different degrees of impaired self-awareness but comparable visuospatial impairment.
Patients and Methods
Sixty-three patients with unilateral right hemispheric stroke (22 females) and 18 healthy control subjects (7 females) gave written informed consent. Lesions were confirmed by CT or MRI. The study had been approved by the local ethics committee.
Participants performed 6 paper-and-pencil tests from the Behavioral Inattention Test5 (star and line cancellation, line bisection, figure copying, clock drawing, text reading). Subsequent to each subtest, the patients rated their performance on a 5-point scale (ranging from 1, severe difficulties to 5, no difficulties). An equivalent rating was performed by one of the investigators (P.E.) to evaluate the patient's difficulties during task performance. This external rating was moreover used to relate the patients' performance to that of healthy control subjects (patients performing >3 SDs below the mean of control subjects were regarded to have substantial visuospatial deficits; Table 1). Interrater agreement with 2 additional raters (evaluating the performance of 39 patients who gave their consent to video recordings) and correlations with objective performance measures (line bisection deviation and cancellation laterality quotients) were determined to ensure the reliability and validity of the external ratings.
To quantify the degree of unawareness for visuospatial neglect, the following index was calculated:
To evaluate the validity of this new measure, the anosognosia index was correlated with the Awareness Questionnaire6 (n=62).
Lesions were mapped on a template brain with 5-mm axial slice distance using MRIcroN. Due to the high correlation between anosognosia and the severity of visuospatial neglect (see the online-only Data Supplement), lesion analyses were confined to specific patient subgroups. For patients showing visuospatial deficits, patient pairs with identical external rating were identified. Subsequently, each pair member was allocated to 1 of 2 groups based on his or her self-rating. Patients without matching pair members (n=5) or with identical scores (n=2) had to be discarded from these analyses. This procedure yielded 2 groups of 11 patients each with identical externally evaluated performance but different degrees of anosognosia. Importantly, all other measures of visuospatial neglect did not differ significantly between both groups (Table 2). Lesions were compared with a voxelwise Liebermeister test7 using a permutation threshold of P<0.05.
Correlation analyses between the demographic and neuropsychological variables revealed that anosognosia was more severe with increasing age (see the online-only Data Supplement). Moreover, the degree of anosognosia significantly correlated with the performance in the 2 cancellation tasks, ipsilesional line bisection errors and with the external rating score.
Lesion analysis was based on matched patient subgroups with identical test performance but different degrees of anosognosia (Table 2).
The Figure depicts the statistical test on the lesion locations of the 2 subgroups. In the group with more severe anosognosia, a significantly higher number of patients had lesions affecting the right angular gyrus and the superior temporal gyrus when compared with the group with lower anosognosia (but comparable visuospatial impairment).
To illustrate the lesion locations related to the presence of visuospatial neglect irrespective of anosognosia in these patients per se, the lesions of these patients were contrasted with those of the remaining patients with normal performance. This analysis revealed that regions within a dorsal frontoparietal network and underlying white matter, together with inferior parietal brain areas (Figure), were more frequently damaged in patients with severely impaired visuospatial functions.
Lesion volume was significantly higher in patients with strong spatial neglect than in the remaining patients (t=4.4; P<0.001), whereas no significant lesion volume difference was observed between the matched higher and lower anosognosia groups (t=−1.3; P=0.226).
The total rating score of all 3 raters revealed high interrater agreement (mean Pearson correlation coefficient r=0.973; all 3 coefficients >0.96; P<0.001) and was highly correlated with the laterality quotients and line bisection errors (see the online-only Data Supplement). The anosognosia index significantly correlated with the Awareness Questionnaire (r=−0.46; P<0.001).
By contrasting subjective with objective performance ratings in visuospatial paper-and-pencil tests, the present study investigated anosognosia for spatial neglect and its neural anatomy. Independent of the severity of the visuospatial deficit per se, damage to the right inferior parietal and superior temporal cortex was associated with higher anosognosia, whereas anosognosia was not related to lesion size. The new anosognosia test showed good reliability and validity.
Anosognosia was strongly associated with different measures of visuospatial neglect. Moreover, as previously reported,8 anosognosia was more severe with higher age. Hence, age might constitute an important factor for the development of anosognosia. Interestingly, unawareness for hemiplegia has also been critically related to visuospatial deficits.9 Note, however, that anosognosia for hemiplegia has almost exclusively been studied in patients with visuospatial impairment.10,11 Although contrasting lesions of patients with spatial neglect with and without anosognosia for hemiplegia controls for the presence of a visuospatial impairment, the severity of the visuospatial impairments may still differ between groups. For this reason, a matching procedure was used in the current study to control for the severity of the visuospatial deficits. This analysis revealed that lesions affecting the right angular and right superior temporal gyrus were more prevalent in the group with more severe anosognosia. This finding is in line with theoretical models of impaired self-awareness for attentional deficits.12 Moreover, our results parallel findings from Berti et al10 in the domain of hemiplegia, because they also suggest function-specific neural correlates of self-awareness.
Patients with parietal lesions fail to integrate information about performance outcome into future task execution plans, that is, they do not benefit from error feedback in subsequent trials of a task.13 Along these lines, anosognosia for spatial neglect may be caused by a deficient performance monitoring in relation to discrepancies between intended and actual achievements. Further support for this notion comes from studies in the motor system, which show that the parietal cortex is crucially involved in the conscious detection of discrepancies between intended actions and perceived movement consequences.14
Sources of Funding
S.V. is supported by the Deutsche Forschungsgemeinschaft (Vo 1733/1-1).
We thank all our patients and are grateful to our colleagues from the Research Centre Jülich and the University Hospital Cologne as well as the Max-Planck Institute for Neurological Research Cologne for providing the opportunity to acquire MRI scans.
The online-only Data Supplement is available with this article at http://stroke.ahajournals.org/lookup/suppl/doi:10.1161/STROKEAHA.112.657288/-/DC1.
- Received March 12, 2012.
- Revision received April 5, 2012.
- Accepted April 6, 2012.
- © 2012 American Heart Association, Inc.
- Prigatano GP
- Orfei MD,
- Robinson RG,
- Prigatano GP,
- Starkstein S,
- Ruesch N,
- Bria P,
- et al
- Azouvi P,
- Samuel C,
- Louis-Dreyfus A,
- Bernati T,
- Bartolomeo P,
- Beis JM,
- et al.
- Wilson BA,
- Cockburn J,
- Halligan P
- Vocat R,
- Staub F,
- Stroppini T,
- Vuilleumier P
- Berti A,
- Bottini G,
- Gandola M,
- Pia L,
- Smania N,
- Stracciari A,
- et al
- Karnath H-O,
- Baier B,
- Nägele T
- Prigatano GP