CASE REPORT
Paradoxical State Anxiety and Working Memory in a Patient with Acute Stroke Vincent Monfort, PhD,* Florent Bernardin, MSc,* Antoine Grosdemange, MSc,* Xavier Ducrocq, MD, PhD,w Philippe Mathieu, MSc,w and Benoıˆt Bolmont, PhD*
Objective: In view of the negative impact of anxiety on working memory, we induced anxiety in 26 patients with acute stroke and 33 healthy controls, and studied how the anxiety affected their emotional reactivity and how the reactivity affected their verbal and visuospatial working memory. We compared the overall findings with those in 1 of our patients (C.B.) who had presented with an abnormally high level of state anxiety. Methods: We gave verbal and visuospatial 1-back tasks under both neutral and anxiogenic conditions, and we compared participants’ working memory scores, self-reported levels of state anxiety, and electrodermal activity. Results: When comparing performance in the neutral condition, the control and patient groups exhibited disrupted verbal working memory, which was associated with greater electrodermal activity and higher state anxiety during the anxiogenic condition. Although patient C.B. also had heightened electrodermal activity during the anxiogenic condition, she experienced a significant reduction in her state anxiety. Her verbal working memory was better during the anxiogenic than the neutral condition. Conclusions: Because of the phonological (subvocal speech) nature of verbal working memory, a higher level of anxious apprehension could explain the increase in state anxiety and the corresponding disruption of verbal working memory in our patient and control groups during the anxiogenic condition. C.B.’s lower state anxiety and selective improvement in verbal working memory during the anxiogenic condition suggest that she felt less anxious apprehension. Key Words: working memory, anxiety, stroke, stress induction (Cogn Behav Neurol 2013;26:195–207)
Reader Benefit: Anxiety can impair cognitive function in people who have suffered brain damage.
Received for publication April 30, 2011; accepted October 30, 2013. From the *Universite´ de Lorraine, LCOMS (EA 7306), Metz, France; and wService de Neurologie, Hoˆpital Central, CHU Nancy, Nancy, France. The authors declare no conflicts of interest. Reprints: Antoine Grosdemange, MSc, Universite´ de Lorraine, LCOMS (EA 7306), Campus Bridoux, Rue du Ge´ne´ral Delestraint, 57070 Metz, France (e-mail: [email protected]). Copyright r 2013 by Lippincott Williams & Wilkins
Cogn Behav Neurol
Volume 26, Number 4, December 2013
ANOVA = analysis of variance; NS = not significant; SCL = skin conductance level; SCW = Stroop-like color word; STAI-Y = State-Trait Anxiety Inventory, Form Y1 (trait anxiety) and Form Y2 (state anxiety); WAIS-R = Wechsler Adult Intelligence Scale–Revised.
A
lthough the negative impact of anxiety on working memory is well-documented in healthy people (Eysenck et al, 2007), little is known about how anxiety contributes to working memory disturbances in braindamaged patients. Numerous studies have suggested that anxiety levels, as measured by subjective state anxiety questionnaires and autonomic responses, may rise because of the anxiogenic effect of neuropsychological tests that require high attention (Hoehn et al, 1997; Silva and Leite, 2000). Anxiety has been shown to affect performance on these tests (Horwitz and McCaffrey, 2008). During the first weeks after a stroke, 25% to 50% of patients have been found to have anxiety (Carota et al, 2002), and 87.6% to have working memory impairment (Jaillard et al, 2009). Regardless of whether the anxiety is related to organic lesions, functional impairments, or psychosocial factors, so many patients with acute stroke are affected that it is relevant to ask whether the disruption of working memory is caused by working memory impairments, the negative impact of anxiety on working memory, or both. The attentional load imposed by anxiety has been widely explored. Numerous studies of healthy people suggest that anxiety primarily affects the central executive component of working memory (Eysenck and Calvo, 1992; Eysenck et al, 2007). Because the central executive component coordinates the function of both the verbal and visuospatial modalities of working memory, anxietyprovoking conditions can disrupt both modalities to similar extents (Eysenck et al, 2007). Other studies have associated anxiety with a selective depletion of phonological (subvocal speech) (Ikeda et al, 1996; Markham and Darke, 1991) or visuospatial (Shackman et al, 2006) resources in working memory. Shackman et al (2006) suggested that the selective disruption of different working memory modalities may result from differential effects of anxious arousal (symptoms of physiologic hyperarousal) and anxious apprehension (worry and verbal rumination). The investigators www.cogbehavneurol.com |
195
Monfort et al
Cogn Behav Neurol
reported that anxious arousal, such as that induced by the threat of an electric shock, had a deleterious effect on participants’ performance of visuospatial tasks. The impairment might be explained by anxious arousal competing with visuospatial working memory for right hemisphere resources (Shackman et al, 2006). Conversely, anxious apprehension and verbal working memory compete for left hemisphere resources (Engels et al, 2007). Impaired verbal performance is expected during academic testing and the test-like conditions used in the research laboratory (in these studies, performance evaluation and failure feedback), and both academic and research testing are mediated by anxious apprehension (Coy et al, 2011; Morris et al, 1981; Shackman et al, 2006). All these studies have shown that anxiety can be induced by difficult neuropsychological tests. Anxiety is also known to have detrimental effects on a broad range of tasks that require working memory. Thus, anxiety must be considered when interpreting working memory performance during the neuropsychological assessment of patients with acute stroke. In this context, we investigated the impact of anxiety on verbal and visuospatial working memory in patients who were in the acute phase of a stroke. We designed a behavioral experiment to compare changes in verbal and visuospatial working memory performance under neutral and anxiogenic conditions in patients in the acute phase of a stroke and in healthy controls. We will report our overall comparison of the patient and control groups in a later paper (manuscript submitted, November 2013). In this article, we report some general results and focus on the unusual findings in 1 of our study patients, a woman whom we are calling “C.B.” C.B. had suffered a stroke in her right middle cerebral artery. Just before C.B. started the behavioral experiment, her self-report showed that she had a pathologic level of baseline state anxiety. During the experiment, she exhibited a unique pattern of emotional reactivity: She had an abnormally high level of state anxiety during the neutral condition, but her anxiety fell during the anxiogenic condition. By contrast, anxiety increased in the other patients and the controls during the anxiogenic condition. Thus, the change in C.B.’s anxiety levels between the neutral and anxiogenic conditions was the opposite of what we observed in the other 58 study participants. C.B.’s results gave us a unique opportunity to study whether changes in state anxiety levels under neutral and anxiogenic conditions can selectively modulate verbal or visual working memory performance, thus adding to the understanding of the relationship between anxiety and working memory in the acute phase of stroke.
seniors’ organizations in Nancy. We performed the study at the University Hospital of Nancy between February 2009 and August 2011. To be included in the study, all of the patients had to meet these inclusion criteria: no previous psychiatric or neurologic history, no change in psychotropic therapies within the previous month (in France, people may be prescribed anxiolytic or antidepressant drugs without having been given an explicit psychiatric or neurologic diagnosis), and no severe cognitive impairments in instrumental or executive function or working memory. To be included in the study, the controls had to have age- and education-adjusted Mini-Mental State Examination (Folstein et al, 1975) and working memory scores higher than the cutoffs, indicating no current dementia or working memory deficits. Each control candidate underwent a structured clinical interview conducted by a senior neurologist, enabling us to exclude volunteers who were undergoing psychotropic treatment or with a history of psychiatric or neurologic disorders. Although we did not test for colorblindness, we asked all participants if they were colorblind and all said that they were not. We matched C.B. and the other participants by age (C.B. 77 years old; other patients 66.6 ± 8.2 years; controls 68.7 ± 9.7 years), hand dominance (C.B. and all other participants were right-handed), and education level (C.B. 9 years; other patients 11.2 ± 2.2 years; controls 11.4 ± 3.3 years). A Crawford and Howell modified t test (see the “Data Analysis” section below) showed no significant difference between C.B. and the other groups in age (patients: t24 = 1.24; P = 0.23; controls: t32 = 0.84; P = 0.40) or education level (patients: t24 = 0.98; P = 0.34; controls: t32 = 0.72; P = 0.48). We gave all participants an initial neuropsychological assessment, followed by the experiment that assessed their anxiety and working memory through verbal and visuospatial 1-back tasks, self-reports of state anxiety, and measurements of electrodermal activity. We gave the patients the neuropsychological evaluation between 5 and 20 days after their stroke, while they were still in the acute phase and still in the hospital; they completed the experiment the following day. The controls did everything in 1 day: First, they had the structured interview and the neuropsychological evaluation; then, after a 30-minute rest period, they completed the experiment. In accordance with French law, our experimental protocol was approved by the local ethical review board, the Comite´ de Protection des Personnes Est III. Patient C.B. and the other participants gave informed consent.
Volume 26, Number 4, December 2013
Patient C.B.’s Hospital Course METHODS Participants C.B. was 1 of 26 patients whom we recruited for our study within the first month after their first-ever stroke. All had been referred to the Stroke Unit of the University Hospital of Nancy. We recruited 33 healthy controls from
196 | www.cogbehavneurol.com
C.B. was a 77-year-old right-handed woman in the acute phase of a cardioembolic stroke. She was admitted to our Stroke Unit after the sudden onset of left hemiplegia. For several years before the stroke, she had been treated for hypertension and hypercholesterolemia; she was also known to have lone atrial fibrillation, but it was not being treated. r
2013 Lippincott Williams & Wilkins
Cogn Behav Neurol
Volume 26, Number 4, December 2013
State Anxiety and Working Memory in Acute Stroke
Initial examination revealed severe hemiparesis, dominant on the left side of her face and left arm. We found no sensory deficits, lateral homonymous hemianopia, speech disturbances, unilateral spatial neglect, or anosognosia. A computed tomogram performed at admission did not reveal any lesions. We diagnosed C.B. as having had an ischemic stroke related to permanent lone atrial fibrillation. We started her on oral anticoagulant therapy, and her left hemiparesis partially improved. She first recovered minor movements of her left arm and hand. Later she was able to walk with help from another person. We gave C.B. a neuropsychological evaluation on day 15 of her hospitalization, during the acute phase of her stroke, and the experimental procedure on day 16. After 3 weeks in the hospital, we discharged C.B. to a rehabilitation center, where she continued to make progress over the next 3 weeks. Because she had lived alone before her stroke and had no relatives nearby, she was released to live in an apartment for senior citizens. Although C.B. gave informed consent for our neuropsychological assessment and behavioral experiment, she did not give consent for further brain imaging.
To seek signs of visual neglect, we gave only the patients the Bells Test (Gauthier et al, 1989), extracted from the Batterie de l’E´valuation de la Ne´gligence and normalized by the Groupe d’E´tude sur la Re´e´ducation et l’E´valuation de la Ne´gligence (GEREN, 2002). Scoring is based on the number of omissions. C.B. again performed normally.
Neuropsychological Evaluation To assess our participants’ cognitive function, we gave them a battery of standardized tests adapted for use with French-speaking people. We used all available normative data that had been age- and education-adjusted for clinical interpretation. Below we describe the tests in the sequence performed. We distinguish the tests given to all participants from those given only to C.B. and the other patients. We did not give all the tests to the controls because we were able to select them based on the structured interview and on tests of their global cognitive efficiency and their working memory. Table 1 shows C.B.’s neuropsychological test scores plus the normative data. Table 2 compares C.B.’s scores to those of the other patients and the controls.
General and Instrumental Function We tested first for global cognitive efficiency. We gave all study participants the Mini-Mental State Examination (Folstein et al, 1975). We also gave the patients (but not the controls) the Mattis Dementia Rating Scale (Schmidt et al, 1994), a rapid test of attention, initiation, construction, conceptualization, and memory. According to both tests, C.B.’s global cognitive efficiency was within the normal range. We tested only the patients for language. We gave the Token Test (De Renzi and Faglioni, 1978), which consists of decontextualized and unusual verbal comprehension tasks, and the Oral Denomination Test DO 80 (Deloche and Hannequin, 1997), which asks participants to speak the names of objects depicted in 80 line drawings. According to these tests, C.B. had no deficits in her language comprehension or speech production. r
2013 Lippincott Williams & Wilkins
Working Memory and Executive Function We assessed working memory efficiency in all of the study participants with the Wechsler Adult Intelligence Scale–Revised (WAIS-R) Digit Span subtests (Wechsler, 1987; Wilde et al, 2004) and with the forward and backward Corsi Block-Tapping Tasks (Wechsler, 1987; Wilde et al, 2004). We gave these tests of working memory to exclude candidates who had severe working memory impairments, to prevent floor effect in our behavioral experiment. C.B.’s score on the WAIS Digit Span Forward showed her to have normal short-term verbal memory. Her short-term visuospatial memory could be considered mildly impaired (Strauss et al, 2006) because her score on the forward Corsi Block-Tapping Task was within 1.5 standard deviations below the mean (Table 1). She had normal scores on the WAIS-R Digit Span Backward for verbal working memory and on the backward Corsi Block-Tapping Task for visuospatial working memory. We tested only the patients for executive function, assessing their mental flexibility with the Trail Making Test (Reitan, 1958), Parts A and B. We based scoring on the total time needed to complete each part of the test and on the number of perseverative errors on Part B, defined as failures to proceed to the next item in the correct order (Ashendorf et al, 2008; Meulemans, 2008). C.B. performed normally on both parts.
Mood We gave all study participants the State-Trait Anxiety Inventory, Form Y1 (STAI-Y1) to report their state anxiety and Form Y2 (STAI-Y2) to report their trait anxiety (Spielberger, 1983). C.B. reported high state anxiety but low trait anxiety. We gave only the patients the Beck Depression Inventory (Beck et al, 1961) and the 10-Item Toronto Alexithymia Rating Scale (Bagby et al, 1994). C.B. was shown to have mild to moderate depression and moderate alexithymia (Taylor et al, 1997). We did not test the controls for depression or alexithymia because our structured clinical interview had enabled us to exclude candidates who had psychiatric disorders.
Behavioral Experiment We investigated how all of our participants’ working memory evolved from a neutral condition to an anxiogenic condition. We asked them to report their state anxiety before the beginning of the experiment (baseline), after the neutral condition, and after the anxiogenic condition. To determine whether the possible increase in their self-reported anxiety during the anxiogenic www.cogbehavneurol.com |
197
Cogn Behav Neurol
Monfort et al
Volume 26, Number 4, December 2013
TABLE 1. Patient C.B.’s Neuropsychological Test Scores in Context Test
C.B.’s Scores
Normative Data or Interpretation for C.B.
Global Cognitive Efficiency Mini-Mental State Examination Mattis Dementia Rating Scale Global Score Attention Initiation Construction Conceptualization Memory
27/30
Cutoff Score 67
State-Trait Anxiety STAI-Y1 (baseline) STAI-Y2
57 37
High state anxiety8 Low trait anxiety8
Depression Beck Depression Inventory
18
Mild to moderate depression9
Alexithymia Toronto Alexithymia Rating Scale
55
Moderate alexithymia10
1, Kalafat et al, 2003. 2, Schmidt et al, 1994. 3, De Renzi and Faglioni, 1978. 4, Deloche and Hannequin, 1997. 5, GEREN, 2002. 6, Wechsler, 1987. 7, Roussel and Godefroy, 2008. 8, Spielberger, 1983. 9, Ve´zina et al, 1990. 10, Loas et al, 2001. SD indicates standard deviation; WAIS-R, Wechsler Adult Intelligence Scale–Revised; STAI-Y1, StateTrait Anxiety Inventory, Form Y1 [1 = state]; STAI-Y2, State-Trait Anxiety Inventory, Form Y2 [2 = trait].
condition was associated with an increase in general arousal, we recorded their autonomic responses during the experiment. We gave the participants 2 tasks, 1 to measure their verbal and visuospatial working memory, and the other just to induce anxiety.
1-Back Working Memory Task The real experimental task was an N-back working memory task. We gave a verbal and a visuospatial 1-back task similar to those described by Shackman et al (2006). We developed and validated the task ourselves. We gave the task the same way in both the neutral and anxiogenic conditions, and we analyzed the scores for both conditions. In each trial of the 1-back task, we showed the participants 1 of 6 possible alphabet letters (b, c, d, h, j, k) on a computer screen. They had to judge whether the current letter matched the letter shown immediately before, and had to use their right hand to press a button for
198 | www.cogbehavneurol.com
“identical” or “different.” Each letter appeared onscreen in a pseudo-random order for 500 milliseconds, followed by a 2500-millisecond blank screen; participants had a total of 3 seconds to see and respond to each letter before the next one appeared. We showed a sequence of 5 letters per trial, asking participants to give 4 1-back responses. At the beginning of each task, we gave the participants onscreen instructions about whether we wanted them to match the actual letter, eg, “b” with “b” (verbal condition), or the letter’s location onscreen, eg, “top right corner” (visuospatial condition). All of the 1-back tasks contained identical stimuli and required identical responses; they differed only in whether the participants had to process the letter’s name or its location. We did not give participants any feedback on their performance.
Stroop-Like Color-Word (SCW) Matching Task We included another task in the experiment just to induce a moderate level of state anxiety during the r
2013 Lippincott Williams & Wilkins
Cogn Behav Neurol
Volume 26, Number 4, December 2013
State Anxiety and Working Memory in Acute Stroke
TABLE 2. Patient C.B.’s Neuropsychological Test Scores Compared to Those of the Other Study Patients and the Controls Study Groups’ Mean (SD) Test
C.B.
Other Patients (n = 25)
P
Controls (n = 33)
Global Cognitive Efficiency Mini-Mental State Examination Mattis Dementia Rating Scale Global Score
27
27.6 (1.7)
NS
134
139.1 (3.3)
NS
—
Language Token Test Oral Denomination Test DO 80
33 79
34.4 (1.7) 78.9 (1.5)
NS NS
— —
Visual Neglect Bells Test
0
1.1 (1.5)
NS
—
Working Memory WAIS-R Digit Span Forward WAIS-R Digit Span Backward Forward Corsi Block-Tapping Task Backward Corsi Block-Tapping Task
5 4 4 4
5.8 5.0 5.8 5.0
(0.9) (1.1) (0.8) (0.8)
NS NS * NS
35 103
43 (18) 143 (60)
NS NS
State-Trait Anxiety STAI-Y1 (baseline) STAI-Y2
57 37
28.7 (8.6) 38.7 (9.8)
** NS
Depression Beck Depression Inventory
18
7.6 (6.0)
NS
—
Alexithymia Toronto Alexithymia Rating Scale
55
48.4 (12.2)
NS
—
Executive Function: Mental Flexibility Trail Making Test Part A (seconds) Trail Making Test Part B (seconds)
29 (1.0)
5.8 4.8 5.5 5.0
(0.9) (1.1) (0.8) (0.6)
P NS
NS NS NS NS
— — 27.8 (7.3) *** 39.7 (8.2) NS
Significant at *P
Paradoxical State Anxiety and Working Memory in a Patient with Acute Stroke Vincent Monfort, PhD,* Florent Bernardin, MSc,* Antoine Grosdemange, MSc,* Xavier Ducrocq, MD, PhD,w Philippe Mathieu, MSc,w and Benoıˆt Bolmont, PhD*
Objective: In view of the negative impact of anxiety on working memory, we induced anxiety in 26 patients with acute stroke and 33 healthy controls, and studied how the anxiety affected their emotional reactivity and how the reactivity affected their verbal and visuospatial working memory. We compared the overall findings with those in 1 of our patients (C.B.) who had presented with an abnormally high level of state anxiety. Methods: We gave verbal and visuospatial 1-back tasks under both neutral and anxiogenic conditions, and we compared participants’ working memory scores, self-reported levels of state anxiety, and electrodermal activity. Results: When comparing performance in the neutral condition, the control and patient groups exhibited disrupted verbal working memory, which was associated with greater electrodermal activity and higher state anxiety during the anxiogenic condition. Although patient C.B. also had heightened electrodermal activity during the anxiogenic condition, she experienced a significant reduction in her state anxiety. Her verbal working memory was better during the anxiogenic than the neutral condition. Conclusions: Because of the phonological (subvocal speech) nature of verbal working memory, a higher level of anxious apprehension could explain the increase in state anxiety and the corresponding disruption of verbal working memory in our patient and control groups during the anxiogenic condition. C.B.’s lower state anxiety and selective improvement in verbal working memory during the anxiogenic condition suggest that she felt less anxious apprehension. Key Words: working memory, anxiety, stroke, stress induction (Cogn Behav Neurol 2013;26:195–207)
Reader Benefit: Anxiety can impair cognitive function in people who have suffered brain damage.
Received for publication April 30, 2011; accepted October 30, 2013. From the *Universite´ de Lorraine, LCOMS (EA 7306), Metz, France; and wService de Neurologie, Hoˆpital Central, CHU Nancy, Nancy, France. The authors declare no conflicts of interest. Reprints: Antoine Grosdemange, MSc, Universite´ de Lorraine, LCOMS (EA 7306), Campus Bridoux, Rue du Ge´ne´ral Delestraint, 57070 Metz, France (e-mail: [email protected]). Copyright r 2013 by Lippincott Williams & Wilkins
Cogn Behav Neurol
Volume 26, Number 4, December 2013
ANOVA = analysis of variance; NS = not significant; SCL = skin conductance level; SCW = Stroop-like color word; STAI-Y = State-Trait Anxiety Inventory, Form Y1 (trait anxiety) and Form Y2 (state anxiety); WAIS-R = Wechsler Adult Intelligence Scale–Revised.
A
lthough the negative impact of anxiety on working memory is well-documented in healthy people (Eysenck et al, 2007), little is known about how anxiety contributes to working memory disturbances in braindamaged patients. Numerous studies have suggested that anxiety levels, as measured by subjective state anxiety questionnaires and autonomic responses, may rise because of the anxiogenic effect of neuropsychological tests that require high attention (Hoehn et al, 1997; Silva and Leite, 2000). Anxiety has been shown to affect performance on these tests (Horwitz and McCaffrey, 2008). During the first weeks after a stroke, 25% to 50% of patients have been found to have anxiety (Carota et al, 2002), and 87.6% to have working memory impairment (Jaillard et al, 2009). Regardless of whether the anxiety is related to organic lesions, functional impairments, or psychosocial factors, so many patients with acute stroke are affected that it is relevant to ask whether the disruption of working memory is caused by working memory impairments, the negative impact of anxiety on working memory, or both. The attentional load imposed by anxiety has been widely explored. Numerous studies of healthy people suggest that anxiety primarily affects the central executive component of working memory (Eysenck and Calvo, 1992; Eysenck et al, 2007). Because the central executive component coordinates the function of both the verbal and visuospatial modalities of working memory, anxietyprovoking conditions can disrupt both modalities to similar extents (Eysenck et al, 2007). Other studies have associated anxiety with a selective depletion of phonological (subvocal speech) (Ikeda et al, 1996; Markham and Darke, 1991) or visuospatial (Shackman et al, 2006) resources in working memory. Shackman et al (2006) suggested that the selective disruption of different working memory modalities may result from differential effects of anxious arousal (symptoms of physiologic hyperarousal) and anxious apprehension (worry and verbal rumination). The investigators www.cogbehavneurol.com |
195
Monfort et al
Cogn Behav Neurol
reported that anxious arousal, such as that induced by the threat of an electric shock, had a deleterious effect on participants’ performance of visuospatial tasks. The impairment might be explained by anxious arousal competing with visuospatial working memory for right hemisphere resources (Shackman et al, 2006). Conversely, anxious apprehension and verbal working memory compete for left hemisphere resources (Engels et al, 2007). Impaired verbal performance is expected during academic testing and the test-like conditions used in the research laboratory (in these studies, performance evaluation and failure feedback), and both academic and research testing are mediated by anxious apprehension (Coy et al, 2011; Morris et al, 1981; Shackman et al, 2006). All these studies have shown that anxiety can be induced by difficult neuropsychological tests. Anxiety is also known to have detrimental effects on a broad range of tasks that require working memory. Thus, anxiety must be considered when interpreting working memory performance during the neuropsychological assessment of patients with acute stroke. In this context, we investigated the impact of anxiety on verbal and visuospatial working memory in patients who were in the acute phase of a stroke. We designed a behavioral experiment to compare changes in verbal and visuospatial working memory performance under neutral and anxiogenic conditions in patients in the acute phase of a stroke and in healthy controls. We will report our overall comparison of the patient and control groups in a later paper (manuscript submitted, November 2013). In this article, we report some general results and focus on the unusual findings in 1 of our study patients, a woman whom we are calling “C.B.” C.B. had suffered a stroke in her right middle cerebral artery. Just before C.B. started the behavioral experiment, her self-report showed that she had a pathologic level of baseline state anxiety. During the experiment, she exhibited a unique pattern of emotional reactivity: She had an abnormally high level of state anxiety during the neutral condition, but her anxiety fell during the anxiogenic condition. By contrast, anxiety increased in the other patients and the controls during the anxiogenic condition. Thus, the change in C.B.’s anxiety levels between the neutral and anxiogenic conditions was the opposite of what we observed in the other 58 study participants. C.B.’s results gave us a unique opportunity to study whether changes in state anxiety levels under neutral and anxiogenic conditions can selectively modulate verbal or visual working memory performance, thus adding to the understanding of the relationship between anxiety and working memory in the acute phase of stroke.
seniors’ organizations in Nancy. We performed the study at the University Hospital of Nancy between February 2009 and August 2011. To be included in the study, all of the patients had to meet these inclusion criteria: no previous psychiatric or neurologic history, no change in psychotropic therapies within the previous month (in France, people may be prescribed anxiolytic or antidepressant drugs without having been given an explicit psychiatric or neurologic diagnosis), and no severe cognitive impairments in instrumental or executive function or working memory. To be included in the study, the controls had to have age- and education-adjusted Mini-Mental State Examination (Folstein et al, 1975) and working memory scores higher than the cutoffs, indicating no current dementia or working memory deficits. Each control candidate underwent a structured clinical interview conducted by a senior neurologist, enabling us to exclude volunteers who were undergoing psychotropic treatment or with a history of psychiatric or neurologic disorders. Although we did not test for colorblindness, we asked all participants if they were colorblind and all said that they were not. We matched C.B. and the other participants by age (C.B. 77 years old; other patients 66.6 ± 8.2 years; controls 68.7 ± 9.7 years), hand dominance (C.B. and all other participants were right-handed), and education level (C.B. 9 years; other patients 11.2 ± 2.2 years; controls 11.4 ± 3.3 years). A Crawford and Howell modified t test (see the “Data Analysis” section below) showed no significant difference between C.B. and the other groups in age (patients: t24 = 1.24; P = 0.23; controls: t32 = 0.84; P = 0.40) or education level (patients: t24 = 0.98; P = 0.34; controls: t32 = 0.72; P = 0.48). We gave all participants an initial neuropsychological assessment, followed by the experiment that assessed their anxiety and working memory through verbal and visuospatial 1-back tasks, self-reports of state anxiety, and measurements of electrodermal activity. We gave the patients the neuropsychological evaluation between 5 and 20 days after their stroke, while they were still in the acute phase and still in the hospital; they completed the experiment the following day. The controls did everything in 1 day: First, they had the structured interview and the neuropsychological evaluation; then, after a 30-minute rest period, they completed the experiment. In accordance with French law, our experimental protocol was approved by the local ethical review board, the Comite´ de Protection des Personnes Est III. Patient C.B. and the other participants gave informed consent.
Volume 26, Number 4, December 2013
Patient C.B.’s Hospital Course METHODS Participants C.B. was 1 of 26 patients whom we recruited for our study within the first month after their first-ever stroke. All had been referred to the Stroke Unit of the University Hospital of Nancy. We recruited 33 healthy controls from
196 | www.cogbehavneurol.com
C.B. was a 77-year-old right-handed woman in the acute phase of a cardioembolic stroke. She was admitted to our Stroke Unit after the sudden onset of left hemiplegia. For several years before the stroke, she had been treated for hypertension and hypercholesterolemia; she was also known to have lone atrial fibrillation, but it was not being treated. r
2013 Lippincott Williams & Wilkins
Cogn Behav Neurol
Volume 26, Number 4, December 2013
State Anxiety and Working Memory in Acute Stroke
Initial examination revealed severe hemiparesis, dominant on the left side of her face and left arm. We found no sensory deficits, lateral homonymous hemianopia, speech disturbances, unilateral spatial neglect, or anosognosia. A computed tomogram performed at admission did not reveal any lesions. We diagnosed C.B. as having had an ischemic stroke related to permanent lone atrial fibrillation. We started her on oral anticoagulant therapy, and her left hemiparesis partially improved. She first recovered minor movements of her left arm and hand. Later she was able to walk with help from another person. We gave C.B. a neuropsychological evaluation on day 15 of her hospitalization, during the acute phase of her stroke, and the experimental procedure on day 16. After 3 weeks in the hospital, we discharged C.B. to a rehabilitation center, where she continued to make progress over the next 3 weeks. Because she had lived alone before her stroke and had no relatives nearby, she was released to live in an apartment for senior citizens. Although C.B. gave informed consent for our neuropsychological assessment and behavioral experiment, she did not give consent for further brain imaging.
To seek signs of visual neglect, we gave only the patients the Bells Test (Gauthier et al, 1989), extracted from the Batterie de l’E´valuation de la Ne´gligence and normalized by the Groupe d’E´tude sur la Re´e´ducation et l’E´valuation de la Ne´gligence (GEREN, 2002). Scoring is based on the number of omissions. C.B. again performed normally.
Neuropsychological Evaluation To assess our participants’ cognitive function, we gave them a battery of standardized tests adapted for use with French-speaking people. We used all available normative data that had been age- and education-adjusted for clinical interpretation. Below we describe the tests in the sequence performed. We distinguish the tests given to all participants from those given only to C.B. and the other patients. We did not give all the tests to the controls because we were able to select them based on the structured interview and on tests of their global cognitive efficiency and their working memory. Table 1 shows C.B.’s neuropsychological test scores plus the normative data. Table 2 compares C.B.’s scores to those of the other patients and the controls.
General and Instrumental Function We tested first for global cognitive efficiency. We gave all study participants the Mini-Mental State Examination (Folstein et al, 1975). We also gave the patients (but not the controls) the Mattis Dementia Rating Scale (Schmidt et al, 1994), a rapid test of attention, initiation, construction, conceptualization, and memory. According to both tests, C.B.’s global cognitive efficiency was within the normal range. We tested only the patients for language. We gave the Token Test (De Renzi and Faglioni, 1978), which consists of decontextualized and unusual verbal comprehension tasks, and the Oral Denomination Test DO 80 (Deloche and Hannequin, 1997), which asks participants to speak the names of objects depicted in 80 line drawings. According to these tests, C.B. had no deficits in her language comprehension or speech production. r
2013 Lippincott Williams & Wilkins
Working Memory and Executive Function We assessed working memory efficiency in all of the study participants with the Wechsler Adult Intelligence Scale–Revised (WAIS-R) Digit Span subtests (Wechsler, 1987; Wilde et al, 2004) and with the forward and backward Corsi Block-Tapping Tasks (Wechsler, 1987; Wilde et al, 2004). We gave these tests of working memory to exclude candidates who had severe working memory impairments, to prevent floor effect in our behavioral experiment. C.B.’s score on the WAIS Digit Span Forward showed her to have normal short-term verbal memory. Her short-term visuospatial memory could be considered mildly impaired (Strauss et al, 2006) because her score on the forward Corsi Block-Tapping Task was within 1.5 standard deviations below the mean (Table 1). She had normal scores on the WAIS-R Digit Span Backward for verbal working memory and on the backward Corsi Block-Tapping Task for visuospatial working memory. We tested only the patients for executive function, assessing their mental flexibility with the Trail Making Test (Reitan, 1958), Parts A and B. We based scoring on the total time needed to complete each part of the test and on the number of perseverative errors on Part B, defined as failures to proceed to the next item in the correct order (Ashendorf et al, 2008; Meulemans, 2008). C.B. performed normally on both parts.
Mood We gave all study participants the State-Trait Anxiety Inventory, Form Y1 (STAI-Y1) to report their state anxiety and Form Y2 (STAI-Y2) to report their trait anxiety (Spielberger, 1983). C.B. reported high state anxiety but low trait anxiety. We gave only the patients the Beck Depression Inventory (Beck et al, 1961) and the 10-Item Toronto Alexithymia Rating Scale (Bagby et al, 1994). C.B. was shown to have mild to moderate depression and moderate alexithymia (Taylor et al, 1997). We did not test the controls for depression or alexithymia because our structured clinical interview had enabled us to exclude candidates who had psychiatric disorders.
Behavioral Experiment We investigated how all of our participants’ working memory evolved from a neutral condition to an anxiogenic condition. We asked them to report their state anxiety before the beginning of the experiment (baseline), after the neutral condition, and after the anxiogenic condition. To determine whether the possible increase in their self-reported anxiety during the anxiogenic www.cogbehavneurol.com |
197
Cogn Behav Neurol
Monfort et al
Volume 26, Number 4, December 2013
TABLE 1. Patient C.B.’s Neuropsychological Test Scores in Context Test
C.B.’s Scores
Normative Data or Interpretation for C.B.
Global Cognitive Efficiency Mini-Mental State Examination Mattis Dementia Rating Scale Global Score Attention Initiation Construction Conceptualization Memory
27/30
Cutoff Score 67
State-Trait Anxiety STAI-Y1 (baseline) STAI-Y2
57 37
High state anxiety8 Low trait anxiety8
Depression Beck Depression Inventory
18
Mild to moderate depression9
Alexithymia Toronto Alexithymia Rating Scale
55
Moderate alexithymia10
1, Kalafat et al, 2003. 2, Schmidt et al, 1994. 3, De Renzi and Faglioni, 1978. 4, Deloche and Hannequin, 1997. 5, GEREN, 2002. 6, Wechsler, 1987. 7, Roussel and Godefroy, 2008. 8, Spielberger, 1983. 9, Ve´zina et al, 1990. 10, Loas et al, 2001. SD indicates standard deviation; WAIS-R, Wechsler Adult Intelligence Scale–Revised; STAI-Y1, StateTrait Anxiety Inventory, Form Y1 [1 = state]; STAI-Y2, State-Trait Anxiety Inventory, Form Y2 [2 = trait].
condition was associated with an increase in general arousal, we recorded their autonomic responses during the experiment. We gave the participants 2 tasks, 1 to measure their verbal and visuospatial working memory, and the other just to induce anxiety.
1-Back Working Memory Task The real experimental task was an N-back working memory task. We gave a verbal and a visuospatial 1-back task similar to those described by Shackman et al (2006). We developed and validated the task ourselves. We gave the task the same way in both the neutral and anxiogenic conditions, and we analyzed the scores for both conditions. In each trial of the 1-back task, we showed the participants 1 of 6 possible alphabet letters (b, c, d, h, j, k) on a computer screen. They had to judge whether the current letter matched the letter shown immediately before, and had to use their right hand to press a button for
198 | www.cogbehavneurol.com
“identical” or “different.” Each letter appeared onscreen in a pseudo-random order for 500 milliseconds, followed by a 2500-millisecond blank screen; participants had a total of 3 seconds to see and respond to each letter before the next one appeared. We showed a sequence of 5 letters per trial, asking participants to give 4 1-back responses. At the beginning of each task, we gave the participants onscreen instructions about whether we wanted them to match the actual letter, eg, “b” with “b” (verbal condition), or the letter’s location onscreen, eg, “top right corner” (visuospatial condition). All of the 1-back tasks contained identical stimuli and required identical responses; they differed only in whether the participants had to process the letter’s name or its location. We did not give participants any feedback on their performance.
Stroop-Like Color-Word (SCW) Matching Task We included another task in the experiment just to induce a moderate level of state anxiety during the r
2013 Lippincott Williams & Wilkins
Cogn Behav Neurol
Volume 26, Number 4, December 2013
State Anxiety and Working Memory in Acute Stroke
TABLE 2. Patient C.B.’s Neuropsychological Test Scores Compared to Those of the Other Study Patients and the Controls Study Groups’ Mean (SD) Test
C.B.
Other Patients (n = 25)
P
Controls (n = 33)
Global Cognitive Efficiency Mini-Mental State Examination Mattis Dementia Rating Scale Global Score
27
27.6 (1.7)
NS
134
139.1 (3.3)
NS
—
Language Token Test Oral Denomination Test DO 80
33 79
34.4 (1.7) 78.9 (1.5)
NS NS
— —
Visual Neglect Bells Test
0
1.1 (1.5)
NS
—
Working Memory WAIS-R Digit Span Forward WAIS-R Digit Span Backward Forward Corsi Block-Tapping Task Backward Corsi Block-Tapping Task
5 4 4 4
5.8 5.0 5.8 5.0
(0.9) (1.1) (0.8) (0.8)
NS NS * NS
35 103
43 (18) 143 (60)
NS NS
State-Trait Anxiety STAI-Y1 (baseline) STAI-Y2
57 37
28.7 (8.6) 38.7 (9.8)
** NS
Depression Beck Depression Inventory
18
7.6 (6.0)
NS
—
Alexithymia Toronto Alexithymia Rating Scale
55
48.4 (12.2)
NS
—
Executive Function: Mental Flexibility Trail Making Test Part A (seconds) Trail Making Test Part B (seconds)
29 (1.0)
5.8 4.8 5.5 5.0
(0.9) (1.1) (0.8) (0.6)
P NS
NS NS NS NS
— — 27.8 (7.3) *** 39.7 (8.2) NS
Significant at *P
