Anosognosia and asomatognosia during intracarotid amobarbital inactivation

Methods.

The subjects were 62 patients undergoing the intracarotid amobarbital procedure as part of their preoperative evaluation for intractable epilepsy. The study was approved by the institutional review board, and subjects gave consent for the procedure. Demographics included: mean age = 34 years (range, 16 to 56); 38 women, 24 men; 57 dextrals, four sinistrals, one ambidextral; 60 with left cerebral language dominance, one with right cerebral language dominance, and one with mixed cerebral language dominance.

Amobarbital was injected by slow hand push into the intracarotid artery following catheterization via a transfemoral approach. The amobarbital dose was 100 mg for each injection in all patients. At the onset of injection, the patients had both hands raised in the air and were counting repetitively from 1 to 20. Eye gaze and language were initially assessed. Subsequent to the testing of anosognosia and asomatognosia described below, memory items were presented, and language was evaluated in greater detail. Further details of our Wada procedure are available elsewhere.15

Approximately 30 seconds (range, 20 to 45) following injection of the non–language-dominant hemisphere, the first 30 patients were asked if either arm felt weak. Then, the hand contralateral to the injection was placed into the visual field ipsilateral to the injection, and the patient was asked, “Whose hand is this?” In the second series of 32 patients, they were initially asked, “Are both your arms in the air?” Next, they were requested point to their left hand with their right hand. Then, they were asked, “Is either one of your arms weak?” and if they replied positively, they were asked, “Which one?” Finally, their left hand was placed in the right visual field, and they were asked, “What is this?” Thus, all patients were asked about weakness and identity of the paralyzed limb, but patients in the second series were also asked about limb location and about limb identity in a more open-ended manner. During the questioning, the arm contralateral to the injection was hemiplegic. Motor strength, anosognosia, and asomatognosia were reassessed in 53 of the patients at a mean of 4.3 minutes (range, 2.9 to 6.7 minutes) postinjection.

Results.

See tables 1 and 2 ⇓ for a summary of data. During the initial postinjection evaluation, five (8%) patients were untestable owing to agitation or depressed level of consciousness. Of the five patients who were initially untestable, three later exhibited anosognosia alone and two had both anosognosia and asomatognosia. Therefore, 88% of the patients exhibited anosognosia and 82% displayed asomatognosia at some point during inactivation of the nondominant cerebral hemisphere (i.e., right brain except for the one patient with right cerebral language dominance). When patients could not recognize their own hand, they uniformly thought that it was someone else’s hand, usually the examiner’s hand. Only two patients (3%) exhibited neither anosognosia nor asomatognosia, and one of these patients had mixed language dominance. During the late follow-up evaluation of 53 patients (2.9 to 6.7 minutes postinjection), the majority of patients had anosognosia alone, and none had asomatognosia alone. Motor strength at this time ranged from 0/5 to 4/5 with a median of 4/5. The distribution of deficits during the early and late evaluations was substantially different (Χ² [3] = 58.53, p < 0.0001). The changes in deficits from early to late evaluations were 79% to 61% for anosognosia and 77% to 8% for asomatognosia. Thus, asomatognosia resolved earlier, decreasing dramatically during this time period, whereas the decrease in anosognosia was modest.

In the second series of patients who underwent additional testing (n = 32), most thought that both arms remained in the air during the acute hemiplegia (table 3). When asked to point to their paralyzed left hand, a minority of patients pointed to the correct position on the table. In the 14 patients who pointed incorrectly, the positions included the original point in the air before injection in seven patients (50%), intermediate positions between the original point and the table in two patients (14%), the shoulder or elbow on the table in two patients (14%), and the right hand in one patient (7%). When asked if either arm was weak, most thought that neither arm was weak. Five (16%) stated that their right arm was weak, but left/right confusion was ruled out by having these patients visually indicate which arm they thought was weak. Thus, a total of 25 of the 32 patients (78%) in this subgroup did not think that their left arm was weak. The majority of patients thought that their left hand was someone else’s hand (the examiner’s hand in all but two cases).

The presence of asomatognosia did not differ between patients with and without anosognosia (Χ² [1] = 0.06, p = 0.80) (table 4). Similarly, the incidence of anosognosia was not affected by the presence of asomatognosia (Χ² [1] = 0.61, p = 0.80) (table 5). Patients with anosognosia were more likely to think that both arms were still in the air (Χ² [1] = 5.83, p < 0.02) and less likely to point to the correct position of the paralyzed hand on the angiography table (Χ² [1] = 3.82, p < 0.05) than patients without anosognosia. However, some patients who denied weakness clearly demonstrated knowledge of the altered location of their affected limb. For example, 12% of the patients with anosognosia knew that both arms were not in the air, and 24% pointed to the correct position of the affected limb on the table.

The presence of asomatognosia was not related to verbal (i.e., answering the question “Are both of your arms in the air?”) (Χ² [1] = 0.72, p = 0.40) or nonverbal (i.e., pointing) (Χ² [1] = 1.22, p = 0.28) knowledge of the paralyzed limb’s location. In fact, nine of the 10 patients who had previously pointed to the correct position of the paralyzed hand had asomatognosia. In contrast, of the eight patients who did recognize their own paralyzed left hand, six (75%) did not know it was weak, and only one (12%) had just been able to point to its correct position on the table. Verbal and nonverbal knowledge of location of the paralyzed limb were unrelated (Χ² [1] = 0.002, p = 0.96) (see table 2).

Discussion.

Anosognosia for hemiplegia and asomatognosia are two of the most dramatic behavioral disorders encountered in clinical neurology. As seen in the current study, both are common with acute inactivation of the non–language-dominant cerebral hemisphere. Only two patients (3%) exhibited neither phenomenon during the Wada test, and one of these patients had atypical cerebral lateralization with mixed left/right language representation. Anosognosia and asomatognosia occurred together in most patients, but they were clearly dissociable. Anosognosia was more common than asomatognosia, especially when the hemispheric anesthesia was less dense later in the Wada procedure. When both phenomena were present initially, asomatognosia resolved first and thus appears to usually require a more widespread or complete dysfunction of the nonlanguage cerebral hemisphere. However, asomatognosia did occur in the absence of anosognosia.

Various mechanisms have been postulated to underlie anosognosia. Psychological denial mechanisms16 based on premorbid personality do not account for the left/right cerebral asymmetries.4 A simple disconnection hypothesis17 seems inadequate given that placing the paralyzed left hand into the right visual field does not alleviate the anosognosia.4 However, performance of the left brain in patients post–corpus callosotomy may not benefit from observing performance of the right brain as evidenced by actions of the left hand.18 Thus, there may be a special role of the right brain in self-awareness.

Confabulations have been postulated to contribute to anosognosia because they occur frequently in stroke patients with anosognosia.19,20 ⇓ Furthermore, illusory limb movements are associated with degree of anosognosia, hemispatial neglect, and asomatognosia.20 In the current study, confabulations (i.e., misperceptions and erroneous attribution) of limb position and body part identity were frequent. Also, they were significantly more common in patients with anosognosia, though they occurred in patients without anosognosia as well. Verbal confabulations appeared to be more common than nonverbal confabulations. During the acute hemiplegia, 72% of the patients stated that both arms were still in the air. This phenomenon could be due to a lack of awareness. However, the patients had been informed before the Wada test to expect weakness, and the time of the injection was indicated aloud during the procedure. Nonverbal confabulations were seen in the 44% of patients who pointed incorrectly to the paralyzed hand. Confabulated locations included the original position in the air, intermediate positions between the original position and the table, the contralateral shoulder, and the contralateral elbow on the table. Confabulations concerning the ownership of the paralyzed hand were seen in 75% of patients. When patients could not recognize their hand, they uniformly thought that it was someone else’s hand. In contrast, the confabulated identity of the affected hand in stroke patients may be more variable in content. Confabulations cannot be the sole cause of denial, but they may play a role in these phenomena by offering an explanation to the conscious mind, which is unaware of the deficits.

A defect in body schema may contribute to anosognosia and asomatognosia.21 Neglect syndrome and inattention to the left hemibody and hemispace probably play an important role. Stroke patients have rarely been reported to have denial of hemiplegia without neglect of the left hemibody.22 In the current study, anosognosia and asomatognosia were frequently concurrent, but they were also dissociable. For example, only 79% of patients with asomatognosia had anosognosia. In addition, 71% of patients without anosognosia had asomatognosia. Such dissociations are much rarer in stroke patients.20 The differences between Wada and stroke patients may be due to differences in acuteness of the lesion, age of the patients, or exam time allocated for the patient to think about their response.

Knowledge of location, weakness, and personal identity of the affected limb were dissociable. Despite believing that both arms were still in the air during the acute hemiplegia, 30% demonstrated a separate knowledge of the location of the paralyzed hand by pointing to the correct position on the table. Verbal knowledge that both arms were not still in the air was present in 12% of patients with anosognosia. Even more patients had nonverbal knowledge of the paralyzed limb’s location: 24% of patients with anosognosia could point to the correct position of the paralyzed hand on the angiography table. In patients who could recognize their own hand, 75% did not know it was weak, 88% thought that both arms were still in the air, and 50% had pointed to an incorrect location of the paralyzed hand. An opposite dissociation was also seen. In the presence of asomatognosia, 21% knew that their arm was weak, 25% knew that both arms were not still in the air, and 38% had previously demonstrated knowledge of location by pointing correctly to the paralyzed hand. In the case of asomatognosia, the knowledge of location was for the position before movement into the opposite visual field. Thus, a lack of proprioceptive feedback might contribute to a patient’s inability to recognize their own hand.

A feed-forward or intentional theory of anosognosia posits that there is a failure of detection due to a loss of motor intention and expectancy of movement.23 Although failure of intention probably contributes to anosognosia, the multiple contrasting dissociations for knowledge of weakness, location, and body part identity seen our study are consistent with the concept that anosognosia may be caused by multiple mechanisms.19,24 ⇓ Thus, normal self-awareness appears to be dependent on several parallel processes.24 In acute hemiplegia, these processes may include inattention to the left hemibody, reduced sensory feedback, impaired intentional mechanisms, a defect in body schema, and a special role of the right brain in self-awareness.