Journal of the International Neuropsychological Society (2012), 18, 334–342. Copyright E INS. Published by Cambridge University Press, 2012. doi:10.1017/S1355617711001780
Free Recall Memory Performance after Aneurysmal Subarachnoid Hemorrhage
Signy Sheldon,1,2 R. Loch Macdonald,1,2,3,4 AND Tom A. Schweizer1,2,3 1Labatt Family Centre of Excellence in Brain Injury and Trauma Research, Keenan Research Centre of the Li Ka Shing Knowledge Institute, St. Michael’s Hospital, Toronto, Canada 2Keenan Research Centre of the Li Ka Shing Knowledge Institute, St. Michael’s Hospital, Toronto, Canada 3Faculty of Medicine, Department of Surgery, University of Toronto, Toronto, Canada 4Division of Neurosurgery, St. Michael’s Hospital, Toronto, Canada
(RECEIVED August 31, 2011; FINAL REVISION November 29, 2011; ACCEPTED November 30, 2011)
Abstract Memory deficits for survivors of aneurysmal subarachnoid hemorrhage (SAH) are common, however, the nature of these deficits is not well understood. In this study, 24 patients with SAH and matched control participants were asked to study six lists containing words from four different categories. For half the lists, the categories were presented together (organized lists). For the remaining lists, the related words were presented randomly to maximize the use of executive processes such as strategy and organization (unorganized lists). Across adjoining lists, there was overlap in the types of categories given, done to promote intrusions. Compared to control participants, SAH patients recalled a similar number of words for the organized lists, but significantly fewer words for the unorganized lists. SAH patients also reported more intrusions than their matched counterparts. Separating patients into anterior communicating artery ruptures (ACoA) and ruptures in other regions, there was a recall deficit only for the unorganized list for those with ACoA ruptures and deficits across both list types for other rupture locations. These results suggest that memory impairment following SAH is likely driven by impairment in the executive components of memory, particularly for those with ACoA ruptures. Such findings may help direct future cognitive-therapeutic programs. (JINS, 2012, 18, 334–342) Keywords: Memory, Short term, Memory disorders, Mental recall, Aneurysm anterior communicating artery, Intracranial aneurysm, Neuropsychological tests
(Larsson et al., 1989, 1994; Ogden, Mee, & Henning, 1993; Toomela et al., 2004). One year post-injury, Ogden et al. (1993) found that patients with SAH still presented with deficits in explicit memory encoding and retrieval. Other reports have indicated recall impairment following SAH with this impairment correlating with a variety of aneurysmal factors, most notably aneurysm location and the volume of blood in the subarachnoid space (Egge et al., 2005; Larsson et al., 1994; Ørbo et al., 2008). These findings, however, are not universal. For example, Fontanella, Perozzo, Ursone, Garbossa, and Bergui (2003) reported no difference between SAH patients and matched control participants on tests of verbal and spatial memory. One reason for these inconsistent findings may be that studies use a diverse range of standardized neuropsychological tests to assess memory. Previous investigations have used measures such as the California Verbal Learning Test (CVLT; Haug et al., 2007), the Rey-Osterrieth Complex Figure– Recall (Diamond, DeLuca, & Kelley, 1997) and subtests of the Wechsler Memory Scale (WMS; Deluca, 1992; Bellebaum et al., 2004). While these standard tests are essential tools for diagnoses
INTRODUCTION Subarachnoid hemorrhage (SAH) is characterized by bleeding into the subarachnoid space with the majority of cases caused by the rupture of cerebral aneurysms (van Gijn & Rinkel, 2001). Many patients recover well, attain good clinical outcome, are functionally independent and free from any physical disabilities, but they often suffer cognitive and emotional deficits that can severely affect their quality of life (Al-Khindi, Macdonald, & Schweizer, 2010; Hackett & Anderson, 2000; Østbye, Levy, & Mayo, 1997). One of the most common cognitive domains affected by SAH is memory (Mayer et al., 2002; Ørbo et al., 2008; Powell, Kitchen, Heslin, & Greenwood, 2004). There are both subjective and objective reports of memory impairments following SAH after the acute stages of recovery Correspondence and reprint requests to: Signy Sheldon, Division of Neurosurgery, Li Ka Shing Knowledge Institute, St. Michael’s Hospital, Room 230, 209 Victoria Street, Toronto, Ontario, M5B 2T8 Canada. E-mail:
[email protected] 334
Memory recall after SAH and for illustrating general cognitive functioning, they rarely allow insight into the precise nature of memory impairment and may overlook more subtle deficits. The goal of the current investigation was to determine more precisely what aspects of memory are impaired in good recovery SAH patients (i.e., those that are functionally independent and reported as neurologically intact) by using an experimental measure that is more sensitive to memory deficits than standardized neuropsychological tests. In particular, we chose to examine explicit memory processes with and without the demand of executive processes given reports of executive impairments after SAH (see Al-Khindi et al., 2010, for a review). Based on this literature, we hypothesize that patients’ with SAH deficit in explicitly encoding or recalling new material is driven by deficits in the organizational processes (executive functions) associated with a memory task. As such, they should be most impaired on memory tests that tax such functions. A second aim of this study was to examine broadly if aneurysm location affects memory performance. While there are some indications that the pathology of aneurysmal SAH are from associated diffuse effects (for example, Ogden et al., 1993), others have suggested that aneurysm location is a key factor in cognitive functioning. Specifically, aneurysms in the anterior communicating artery (ACoA) affect memory more prominently than aneurysms in other locations (Haug et al., 2009). Indeed, such ruptures are often characterized by alterations in cognition and behavior (Deluca & Diamond, 1995; Parkin, Yeomans, & Bindschaedler, 1994). For example, an early study by Deluca (1992) compared the performance of patients with ACoA aneurysms to patients with aneurysms in other locations and found that the ACoA group performed significantly worse on a delayed memory test as well as an executive function test (Wisconsin Card Sorting Test), but performed better on tests of immediate recall, attention and visuo-spatial functions. Haug and colleagues (2009) reported that ACoA aneurysm ruptures result in poorer executive and memory function compared to middle cerebral artery aneurysm ruptures. Manning, Pierot, and Dufour (2005) tested patients with ACoA ruptures and those with non-ACoA ruptures (middle cerebral artery or posterior communicating artery), all treated with coils, on a battery of neuropsychological tests, and found that only 4 of the 24 tests reported any group differences. One test that was more impaired in patients with ACoA aneurysm ruptures was a free recall word list learning task—a task that often requires organization and strategic retrieval (Turner, Cipolotti, Yousry, & Shallice, 2007). In contrast to these results, Tidswell, Dias, Sagar, Mayes, and Battersby (1995) reported no variation in cognitive performance by aneurysm location when testing patients with SAH, even when testing patients that were treated with minimal surgical intervention (i.e., microvascular clips). Together, the reviewed literature suggests a more specific deficit for patients with ruptured ACoA aneurysms, which may be the result of specific damage to the orbital prefrontal regions (Deluca & Diamond, 1995; Parkin et al., 1994) or a disconnection between circuits in the brain that support such additional strategic processes needed for recall or delayed
335 memory tasks (Mavaddat, Kirkpatrick, Rogers, & Sahakian, 2000). Following such reports, we hypothesize that while SAH will affect memory recall generally, the pattern of impairment will be different for patients with aneurysm ruptures in different locations, particularly those with ruptures in the ACoA and those with ruptures in other locations (herein called non-anterior locations). Patients with ACoA aneurysm ruptures (herein anterior aneurysms) will be more impaired on tests that emphasize the organizational or strategic components of memory, given the link between these processes and frontal lobe functioning (regions affected by ACoA ruptures), than on tests that do not emphasize such components. To achieve the goals of our study, we used a free recall test based on a paradigm used by Turner and colleagues (2007). Participants were presented with word lists under two conditions, one for which words from the same category were presented together or blocked in an organized manner and another in which words from the same category were presented in a random or unorganized manner. If mnemonic deficits from SAH are specific to the implementation of executive strategic processes, then the organized presentation of the words should aid in recall of materials compared to the unorganized presented word lists. This should be most pronounced for those with ACoA ruptures.
METHOD Participants Twenty-four patients with surgically treated SAH (the ruptured aneurysm was treated by neurosurgical clipping or endovascular coiling, Table 1) were recruited from the neurovascular clinic at St. Michael’s Hospital. The mean age of the patients was 57.6 years (SE 5 1.5; range 5 37 to 70 years). Thirteen of the patients were female. All patients had at least ten years of education. Twelve of the patients had ruptures in the anterior communicating artery (ACoA) and 12 had ruptures elsewhere, prominently the middle cerebral artery (MCA; 8) but also the posterior communicating artery (PCoA; 4) and basilar tip (2). The mean time since injury was 55.9 months (SE 5 18; range 5 15 to 459 months). Patients were grouped as either anterior or non-anterior patients for the purpose of this study. Patients with aneurysms located at or near the ACoA were classified as anterior and aneurysms anywhere posterior to this region were classified as non-anterior. The justification for this grouping is that ACoA ruptures typically inject blood in more anterior regions than MCA, PCoA ruptures and ruptures in the posterior circulation (basilar tip aneurysms). This classification allowed us to look at the effect of blood in anterior versus non-anterior regions of the brain, an important analysis given the importance of anterior/frontal brain regions to particular memory functions. No patient was clinically amnesic or exhibited gross cognitive disabilities: only five of the 24 participants did not return to work after their aneurysm ruptured. The remaining patients were working part-time, full-time or retired. All patients had a computed tomography (CT) scan or a magnetic
336
S. Sheldon et al.
Table 1. Characteristics of the 24 SAH patient participants
Group Anterior Anterior Anterior Anterior Anterior Anterior Anterior Anterior Anterior Anterior Anterior Anterior Non-anterior Non-anterior Non-anterior Non-anterior Non-anterior Non-anterior Non-anterior Non-anterior Non-anterior Non-anterior Non-anterior Non-anterior
Aneurysm Location
Intervention (months)
Time since incident
ACoA ACoA ACoA ACoA ACoA ACoA ACoA ACoA ACoA ACoA ACoA ACoA PCoA L MCA PCoA PCoA R MCA L MCA L MCA Basilar Basilar PCoA L MCA L MCA
Coil Coil Coil Coil Coil Coil Coil Coil Coil Coil Coil Coil Clip Clip Coil Coil Clip Clip Clip Coil Coil Coil Clip Clip
30 25 37 35.5 39 45.5 83.5 31 30 29.5 31 86.5 42.5 89 15.5 15 41.5 20.5 60 459 21 34 23 19
Battig and Montague (1969) norms (Van Overschelde, Rawson, & Dunlosky, 2004; Table 2). When selecting the words, we excluded the most frequent associates (the first, second, and third) to ensure that one could not rely on prototypical responses and semantic memory. An analysis of variance (ANOVA) revealed no difference between lists (F(1,90) 5 1.008; p 5 .42) in the average normative proportion that the particular word was used in response to its category (Van Overschelde et al., 2004). Words for the first, third, and fifth list were presented blocked: that is, words from each category were presented together to minimize the need to implement any organizational processes or strategy to encode or retrieve the items. The words in the second, fourth, and sixth list were presented in a random manner so to maximize the need to use organization or strategy to encode or retrieve the items. Deviating slightly from Turner et al. (2007), for list two, four, and six, two of the categories used were ones that appeared in the previous list (list one, three, and five, respectively). There was no repetition of words from one list to the other list; rather we used words from the same category to promote intrusions from prior lists and proactive inference (Table 2).
Procedure
resonance imaging (MRI) scan following their admission to the hospital. For each patient participant, the appearance of the hemorrhage was classified according to the modified Fisher Grade based on their admission scan (1 – not evident, 2 – less than 1 mm thick, 3 – more than 1 mm thick, 4 – intraventricular hemorrhage or parenchymal extension; Fisher, Kistler, & Davis, 1980; average 5 3.47; SE 5 0.2). The severity of the subarachnoid hemorrhage was also graded according to the Hunt and Hess (1968) classification (average 5 2.5; SE 5 0.2). Twenty-two healthy control participants that were matched for age and education level were recruited from the community via posters and online advertisements (see Table 3 for demographics). Control participants were free from neurological or psychiatric illness. Participants’ data included in this manuscript was obtained in compliance with the Research Ethics Board at St. Michael’s Hospital. Participants received an honorarium for their participation and all participants gave informed consent. Furthermore, all participants were given a short battery of standardized neuropsychological tests (Strauss, Sherman & Spreen, 2006; Table 3).
For each list, participants were presented with words, one at a time, on the center of the computer screen, for 2 s with a 1-s fixation cross in between each word. Before each list, participants were told to memorize the words on the screen for later recall, but were not told the number of words that they would be given. Directly following the presentation of the last word of the list, a screen came up on the computer that asked participants to count down out loud backward by 3 s from a randomly chosen number that appeared on the screen. They were told to continue to do so until the present screen disappeared. After 30 s of counting backward, the participants were then asked to recall as many words as they could from the previously presented list. They were given as much time as needed. The experimenter recorded the responses. This was repeated for all six lists. After the task was completed, the experimenter classified each word recalled as either correct, a prior list intrusion (words that were falsely recalled from a prior list: these were not possible for list one) or as a semantic or other intrusions (words that were falsely recalled but did not appear on a previous list; most often these words were categorically related to words appearing on the list). The number of words recalled across all three organized lists and all three unorganized lists was tallied. The number of prior list intrusions and semantic/other intrusions was tallied across all lists for each participant.
Stimuli
Statistical Methods
Similar to Turner et al. (2007), six 16-word lists were used in this study. Each list contained four words that belonged to one of four categories taken from an updated version of the
We first compared the performance on standard neuropsychological tests between the groups (patients with SAH and control participants) using independent t tests.
Memory recall after SAH
337
Table 2. Stimuli used for the free recall test List 1
List 2
List 3
List 4
List 5
List 6
flannel satin linen suede plate strainer ladle blender ottoman sofa loveseat futon shoulder ankle torso wrist
jean velvet cashmere fleece bowl whisk stove tongs violet brown orange pink article flyer letter essay
mango kiwi plum peach champagne scotch tequila whiskey scientist engineer carpenter banker beach cliff canyon island
lime melon cherry lemon wine gin martini rum cadet admiral colonel chief tent hotel trailer cabin
hockey running skiing lacrosse coat dress glove belt finch raven hawk canary oregano basil paprika parsley
tennis golf wrestling rugby skirt blouse sweater scarf pig bear cow mouse pastor monk rabbi Minister
Note. All words were taken from categories from the updated Battig and Montague corpus (Reprinted from the Journal of Memory and Language, 50, Van Overschelde, J.P., Rawson, K.A., & Dunlosky, J., Category norms: An updated and expanded versions of the norms, 289–335, 2004, with permission from Elsevier). Boxed letters denote overlap in categorical words across lists.
To examine the results of the experimental memory task, we performed a set of planned between-subject comparisons using ANOVA. Significance was set at p 5 .05. We compared the number of words recalled across all lists and then for the organized lists and the unorganized lists between patients with SAH and control participants. Patient participants were then divided into those with anterior aneurysms and those with ruptures elsewhere (non-anterior aneurysms) and the number of words recalled in the organized and unorganized list conditions was compared. These analyses were repeated with category recall. To compare the benefit of list type between patients with different aneurysm locations to control participants, we looked at the individual difference in the number of words recalled for the organized and unorganized lists (done to control for overall differences in words recall). Similar to Turner et al. (2007), we examined the use of organizational strategy when recalling/retrieving words from the lists. To do so, we calculated the number of switches between categories that were made during retrieval. This was done to assess the number of unnecessary switches using the formula from Turner et al. (2007): (# categories switches made - # categories recalled 1 1) / (# categories recalled 21). For example, if four categories were recalled, then one only needs to switch three times between categories during recall if using an optimal category clustering strategy. Any additional switches are deemed unnecessary and may represent a more disorganized retrieval strategy. We also compared the number of intrusions (words that were falsely recalled during list retrieval) between SAH patients and controls, both those from previous lists and those not from a previous list. Finally, the relation between patient factors, such as a surgical intervention, Hunt and Hess score and Fisher Grade and the words recalled was examined using ANOVA and
other factors, such as time since injury, was examined using correlational analyses.
RESULTS Neuropsychological Tests The mean performance of each group (patients with SAH and control participants) on a subset of standardized neuropsychological tests is reported in Table 3. There was a difference in the estimated premorbid IQ based on the National Adult Reading Test (NART) between healthy controls and patients (F(1,45) 5 4.961; p , .05). Given this result, when comparing the patient and the control participant groups for the experimental task, we included estimated IQ scores as a covariate. There were no differences between the groups in the remaining neuropsychological tests (Table 3).
Free Recall Test Number of recalled words Overall, patients with SAH recalled fewer words compared to control participants (F(2,43) 5 4.017; p , .05). Breaking the lists down to those that were organized (categories of words presented in a blocked manner) to those that were unorganized (categories of words presented randomly throughout the list), patients with SAH recalled significantly fewer items for the unorganized lists (F(2,43) 5 6.610; p , .005), but did not differ significantly from the control participants in terms of the number of words recalled for the organized lists (F(2,43) 5 2.079; p . .10; Figure 1). We also examined how aneurysm location affected memory impairment by comparing patients with anterior (ACoA)
338
S. Sheldon et al. Table 3. Demographic characteristics and mean scores (standard errors shown in the parentheses) on neuropsychological tests for SAH patients and healthy control participants.
Demographic Age Education Number of females Neuropsychological tests NART est IQ CVLT trial 1 CVLT learning (trial 1 to 5) CVLT delay Digit Span Forward Digit Span Backward Digit Symbol (120 seconds) Trails A (seconds) Trails B (seconds)
SAH Patients
Healthy controls
p value (unadjusted)
57.6 (1.5) 15.4 (0.6) 13
60.5 (2.4) 15.0 (0.5) 12
.30 .60
107 (2.0) 5.5 (0.4) 48.1 (1.7) 10.5 (0.5) 9.3 (0.4) 6.7 (0.4) 57.2 (2.9) 31.6 (3.2) 71.4 (4.7)
113(0.9) 6.7 (0.5) 50.5 (2.7) 10.8 (0.8) 10.1 (0.4) 7.2 (0.4) 66.0 (3.6) 31.1 (3.1) 64.5 (5.1)
.03* .07 .45 .76 .14 .42 .14 .97 .61
and those with non-anterior (all others) aneurysms. The patient groups did not differ in terms of sex, age, education level (Table 4), Hunt and Hess Scores (p 5 .13) nor Fisher grade scores (p 5 .86), thus any difference we find are likely due to aneurysm location. Given that there was no difference in estimated IQ based on the NART between these two patient groups, we did not need to include IQ as a covariate for this analysis. Patients with anterior aneurysms recalled more words for the organized lists compared to those with aneurysms in the nonanterior region (F(1,23) 5 7.013; p 5 .015), but a similar number of words for the unorganized lists (F(1,23) 5 1.845; p 5 .19; Figure 2). Examining the benefit of list type between patients with different aneurysm locations to control participants, when we compared anterior aneurysm patients and healthy controls,
there was a significant difference (F(2,33) 5 5.825; p , .005; Figure 3) but this was not the case for the non-anterior aneurysm patients compared to healthy controls (F(2,31) 5 0.690; p 5 .50; Figure 3). This suggests that patients with anterior aneurysms benefited more from the presence of organization in the lists than healthy control participants and non-anterior aneurysm patients. Further supporting this is the finding that non-anterior patients recalled fewer words for both the organized lists (F(2,31) 5 5.490; p , .01) as well as the unorganized lists (F(2,31) 5 6.184; p 5 .005) compared to controls, but anterior aneurysm patients only differed from controls for the unorganized lists (unorganized lists: F(2,33) 5 3.003; p 5 .06), organized lists: F(2,33) 5 0.432; p 5 .65).
Categorical recall When we examined the number of categories recalled for all patients with SAH and control participants, the same pattern described above emerged. On average, fewer categories were recalled for the SAH patients when they were presented in an unorganized manner (patients mean 5 2.6 (SE 5 0.1); controls mean 5 3.2 (SE 5 0.1); F(1,43) 5 7.939; p 5 .007) but a similar number of categories were recalled when presented in an organized, blocked manner (patients mean 5 2.5 (SE 5 0.2); controls mean 5 2.9 (SE 5 0.1); F(1,43) 5 3.586; p 5 .07) compared to healthy controls.
Organizational strategy at retrieval
Fig. 1. The average number of words recalled for the organized and unorganized lists (maximum score 5 48 for each condition) for patients with SAH and healthy control participants (standard error bars are shown).
We examined the use of organizational strategy when recalling/retrieving words from the lists by assessing the number of unnecessary switches (see methods). When we calculated the number of unnecessary switches, we found no significant difference between SAH patients and controls made for organized and unorganized lists (patient mean 5 2.2, SE 5 0.6; control mean 51.7, SE 5 0.3; F(1,45) 5 0.670; p 5 .42).
Memory recall after SAH
339
Table 4. Demographic characteristics and mean scores (standard errors shown in the parentheses) on neuropsychological tests for SAH patients with anterior ruptures and those with non-anterior ruptures Anterior Demographic Age Education Number of females Neuropsychological tests NART est IQ CVLT trial 1 CVLT learning (trial 1 to 5) CVLT delay Digit Span Forward Digit Span Backward Digit Symbol (120 seconds) Trails A (seconds) Trails B (seconds)
Non-anterior
p value (unadjusted)
58.1 (12.7) 15.1 (0.8) 7
57.1 (1.5) 14.8 (0.8) 9
.74 .28
108 (2.8) 5.8 (0.6) 47.5 (2.2) 11.1 (0.8) 9.4 (0.6) 6.8 (0.5) 61.6 (2.0) 29.2 (3.6) 68.9 (7.3)
106(3.2) 5.2 (0.5) 48.8 (2.8) 9.9 (0.6) 9.1 (0.7) 6.5 (0.8) 52.5 (5.2) 34.6 (4.9) 74.4 (10.2)
.64 .39 .66 .33 .71 .92 .20 .51 .63
Intrusions Patients with SAH erroneously recalled a mean of 2.9 (SE 5 0.6) words from a prior list and 2.0 (SE 5 0.5) other words across the whole task compared to a mean of 1.3 (SE 5 0.3) prior list and 0.9 (SE 5 0.2) other word intrusions that were recalled by the control participants (F(2,43) 5 3.369; p , .05, controlled for NART estimated IQ scores). For the patients, there were intrusions for both organized and unorganized lists, but there were significantly more present for the unorganized lists (t(23) 5 3.153; p 5 .004), a difference that was not significant for the control participants (t(21) 5 0.839; p 5 .41).
Patient factors: Clip versus coil and neuropsychological measures Patients treated with clipping or coiling were not significantly different in the number of words recalled (F(1,23) 5 0.156; p 5 .70, F(1,23) 5 0.014; p . .90; F(1,23) 5 0.093; p 5 .80 for
Fig. 2. The average number of words recalled for the organized and unorganized lists (maximum score 5 48 for each condition) for patients with SAH with aneurysms located in the anterior region of the brain and those with aneurysm located in non-anterior regions of the brain (standard error bars are shown).
words recalled in the organized lists, unorganized lists and total recalled, respectively) nor in the number of intrusions made during recall (F(1,23) 5 0.057; p 5 .80). There also was no correlation between the number of words recalled and the time since injury nor was there any significant correlation between free recall scores and Fisher grade or Hunt and Hess scores. In terms of neuropsychological measures, correlations between the CVLT and the experimental free recall task revealed that in patients, the number of words recalled in the first trial of the CVLT correlated highly with the number of words recalled in the free recall task (r 5 0.615; p 5 0.002). Learning during the CVLT (sum of trial 1 to 5) or delayed CVLT recall scores did not correlate with free recall (r 5 0.255; p . .20; r 5 0.147; p . .50, respectively).
DISCUSSION The current investigation confirms that SAH patients with good clinical recovery and who are functionally independent
Fig. 3. The average individual difference score (the number of words recalled in the organized list—the number of words recalled in the unorganized lists) for patients with SAH with aneurysms located in the anterior region of the brain, those with aneurysm located in non-anterior regions of the brain and healthy control participants (standard error bars are shown).
340 have long-term deficits in explicit memory. Even though the tested patients with SAH performed similar to a group of age and education matched healthy controls on a series of standard neuropsychological tests, they were impaired at word recall when given a more sensitive, experimental measure of free recall. While these results are in line with previous reports of memory deficits following SAH (Chahal, Barker-Collo, & Feigin, 2011; Ogden et al., 1993; Toomela et al., 2004), our results go further to suggest a specific locus of impairment: Patients with SAH are impaired more when recalling words under conditions that required additional strategic and organizational processes (unorganized lists) and not under conditions for which such additional processes are not required (organized lists). From this, we suggest that SAH predominately affects executive processes associated with explicit memory. To determine if these impaired executive processes associated with SAH are due to deficits at encoding or retrieval, we looked at organization strategy at retrieval. This was done by looking at the number of switches made between categorical recall—the more unnecessary switches between categories during retrieval is indicative of poorer strategic recall. We did not find a significant difference between SAH patients and control participants in the number of unnecessary switches. Speculating from this, these data suggest that, given that the reported impairments result not from executive deficits at recall, they likely result from less than optimal encoding strategies. That is, impaired strategic or organizational processes during encoding. Our study also found that patients with SAH reported more intrusions during the free recall test than their matched counterparts. Such findings are in line with recent studies that have found an enhanced susceptibility to some types of false memories following ACoA ruptures (Borsutzky, Fujiwara, Brand, & Markowitsch, 2010), but our results extends this executive dysfunction to aneurysmal SAH in general. The increased number of prior word intrusions in patients with SAH is also evidence of proactive interference, an increased susceptibility to prior information (e.g., Shimamura, 1994). Indeed it could be SAH results in poor proactive interference resolution thought supported by frontal regions that is also the cause of the deficit in recall: a poorer ability to inhibit information may lead to poorer encoding of information. Another possibility is that the increased likelihood of those with SAH to falsely report words during recall, particularly those that are from similar categories, may be the result of more gist based encoding (i.e., encoding words on a broader level).
Neuropsychological Tests While the result of our study clearly report memory impairment in patients with SAH on an experimental memory task, such impairments were not detected by standard neuropsychological instruments. Patients with SAH did not differ from controls on a series of neuropsychological tests (notably, the California Verbal Learning Task; CVLT) nor was there a relation between performance on these standard tasks and
S. Sheldon et al. that of the experimental memory measure. While this may be surprising, we take this as evidence that standard tests are likely not sensitive enough to detect subtle memory impairment nor do these tests allow for a more fine-grain examination of affected cognitive processes in populations such as those who have experienced a SAH.
Anterior Versus Non-anterior Aneurysms Another aim of this investigation was to look at the role of aneurysm location on memory recall. Patients were broadly classified as those with ACoA aneurysm ruptures (anterior) and those with aneurysm ruptures in other locations (non-anterior aneurysms). Using this classification, patients with anterior aneurysm ruptures were impaired only on the free recall task when it stressed organizational and strategic possessing (unorganized lists), but not on the free recall task when it did not require these additional processes (organized lists). Conversely, patients with non-anterior aneurysms were impaired on both lists. These findings are line with reports that ACoA aneurysm ruptures affect executive aspects of memory, such as those needed for recall, but spares other aspects of memory, such as those needed for tests of recognition or organized retrieval (Diamond et al., 1997; Simard, Rouleau, Brosseau, Laframboise, & Bojanowsky, 2003). In terms of neurological function, ACoA aneurysm ruptures often result in direct damage to the orbitofrontal cortex and/or results in disconnection in fronto-temporal circuits (Mavaddat et al., 2000), which can result in deficits in memory and personality changes (Alexander & Freedman, 1984; Wright, Boeve, & Malec, 1999). Given that damage from ACoA ruptures is not to the anatomical substrate of mnemonic functioning (the medial temporal lobes) but is that of executive processes, it is expected (and found) that these ruptures will result in more specific impairment on memory tests that require organizational and strategic processing (Bottger, Prosiegel, Steiger, & Yassouridis, 1998; Diamond et al., 1997). Why patients with non-anterior aneurysms (aneurysms that inject blood in more posterior cortical regions than ACoA ruptures) are generally more impaired on this free recall task requires further investigation. One possible reason is that there were more patients treated with surgical clips in the non-anterior patient group, which could have resulted in local brain damage that did not affect those treated with coils (Bellebaum et al., 2004). However others have found no effect of surgical intervention on cognitive performance (for example, Frazer, Ahuja, Watkins, & Cipolotti, 2007). Other possible mechanisms are the focal presence of the subarachnoid blood and related toxic product (Hadjivassiliou et al., 2001) causing focal damage and/or circulatory arrest from SAH resulting in an environment that damages related watershed areas to the affected vasculature (Grote & Hassler, 1988). Unlike subarachnoid blood in anterior regions, affecting regions secondary to memory functioning, subarachnoid blood in regions near non-anterior aneurysm ruptures may have an effect on general memory functioning (e.g., Tariq et al., 2010), perhaps damaging regions of the medial temporal lobes, namely
Memory recall after SAH the hippocampus (Bendel et al., 2006). Indeed, future investigations are needed to clarify the underlying cause of the presented impairment. In summary, while the cause of memory deficits is likely diffuse and rests upon the interaction of many factors (Kreiter et al., 2002), we found evidence that one factor that affects the pattern of memory impairment is aneurysm location. Our results suggest a more general deficit in explicit, episodic memory in non-anterior ruptured aneurysm patients and one of strategic encoding in anterior ruptured aneurysm patients (also see Manning et al., 2005).
GENERAL IMPLICATIONS The presented results have implications for how patients who have experienced SAH return to daily living. All patients tested were functionally independent, but may also have a reduced quality of life that could result from memory impairments. Relating the current results to subjective experiences by a SAH patient, one may have more trouble recalling the specifics of a task, especially if information is not presented in an organized manner. Another problem for people after SAH, as illustrated by our findings, could be with competing information and placing events in context, as noted by their increased susceptibility to intruding information. It is often the case that one is presented with interfering information in daily living, such as trying to remember a grocery list while planning when to pick up the dry cleaning, all while listening to a radio program. While speculative, it may be that these memory processes associated with daily living are affected by SAH. In terms of rehabilitation practices, these results suggest a program that focuses on executive processes and how such processes overlap and interact with memory functioning would be most effective. An example of such a program is Goal Management Training which can help with poor organization use in memory (Levine et al., 2000).
STUDY LIMITATIONS A methodological limitation of the present experiment is the exclusion of more sensitive tests of executive functions and detailed quality of life measures. Future investigations may wish to include such measures to, first, determine if performance on these tests is impaired in patients with SAH and, second, to see if impairment on executive function and assessments of quality of life relates to performance on the experimental measure of memory used in this study. Another limitation of the present study was our small sample size. A goal of future research is to expand on the number of patients with SAH tested so that a more thorough examination of the effect of aneurysm location on memory performance can be done.
CONCLUSION While our results are of interest for understanding how neural regions support memory processes, we also believe that our findings have important implications for understanding
341 deficits following SAH and for determining future cognitive rehabilitation programs. From these results, we speculate that rehabilitation programs that emphasize strategic organization of to-be-remembered information and interference resolution may be successful.
ACKNOWLEDGMENTS The information in this manuscript nor this manuscript has not been previously published. The authors would like to thank Andrea Constantinof and Rebecca Metcalfe for help with testing and scoring. This work was supported by funds from the Labatt Family Centre of Excellence and a Personnel Award from the Heart and Stroke Foundation of Canada awarded to T.A.S. No financial or other conflicts of interest affect this manuscript.
REFERENCES Alexander, M., & Freedman, M. (1984). Amnesia after anterior communicating artery aneurysm rupture. Neurology, 34, 752–757. Al-Khindi, T., Macdonald, R.L., & Schweizer, T.A. (2010). Cognitive and functional outcome after aneurysmal subarachnoid hemorrhage. Stroke, 41, 519–536. Bellebaum, C., Scha¨fers, L., Schoch, B., Wanke, I., Stolke, D., Forsting, M., & Daum, I. (2004). Clipping versus coiling: Neuropsychological follow up after aneurysmal subarachnoid haemorrhage (SAH). Journal of Clinical and Experimental Neuropsychology, 26, 1081–1092. Bendel, P., Koivisto, T., Ha¨nninen, T., Kolehmainen, A., Ko¨no¨nen, M., Hurskainen, H., y Vanninen, R. (2006). Subarachnoid hemorrhage is followed by temporomesial volume loss: MRI volumetric study. Neurology, 67(4), 575–582. Borsutzky, S., Fujiwara, E., Brand, M., & Markowitsch, H. (2010). Susceptibility to false memories in patients with ACoA aneurysm. Neuropsychologia, 48(10), 2811–2823. Bottger, S., Prosiegel, M., Steiger, H., & Yassouridis, A. (1998). Neurobehavioural disturbances, rehabilitation outcome, and lesion site in patients after rupture and repair of anterior communicating artery aneurysm. Journal of Neurology, Neurosurgery, & Psychiatry, 65(1), 93–102. Chahal, N., Barker-Collo, S., & Feigin, V. (2011). Cognitive and functional outcomes of 5-year subarachnoid haemorrhage survivors: Comparison to matched healthy controls. Neuroepidemiology, 37(1), 31–38. Deluca, J. (1992). Cognitive dysfunction after aneurysms of the anterior communicating artery. Journal of Clinical and Experimental Neuropsychology, 14, 924–934. Deluca, J., & Diamond, B.J. (1995). Aneurysm of the anterior communicating artery: A review of neuroanatomical and neuropsychological sequelae. Journal of Clinical and Experimental Neuropsychology, 17, 100–121. Diamond, B.J., Deluca, J., & Kelley, S.M. (1997). Memory and executive functions in amnesic and non-amnesic patients with aneurysms of the anterior communicating artery. Brain, 120, 1015–1025. Egge, A., Waterloo, K., Sjøholm, H., Ingebrigtsen, T., Forsdahl, S., Jacobsen, E.A., & Romner, B. (2005). Outcome 1 year after aneurismal subarachnoid hemorrhage: Relation between cognitive performance and neuroimaging. Acta Neurologica Scandinavica, 112, 76–80.
342 Fisher, C., Kistler, J., & Davis, J. (1980). Relation of cerebral vasospasm to subarachnoid hemorrhage visualized by computerized tomographic scanning. Neurosurgery, 6(1), 1–9. Fontanella, M., Perozzo, P., Ursone, R., Garbossa, D., & Bergui, M. (2003). Neuropsychological assessment after microsurgical clipping or endovascular treatment for anterior communicating artery aneurysm. Acta Neurochirurgica, 145, 867–872. Frazer, D., Ahuja, A., Watkins, L., & Cipolotti, L. (2007). Coiling versus clipping for the treatment of aneurysmal subarachnoid hemorrhage: A longitudinal investigation into cognitive outcome. Neurosurgery, 60(3), 434–441. Hackett, M.L., & Anderson, C.S. (2000). Health outcomes 1 year after subarachnoid hemorrhage: An international populationbased study. Neurology, 55, 658–662. Hadjivassiliou, M., Tooth, C.L., Romanowski, C.A., Byrne, J., Battersby, R.D., Oxbury, S., y Sagar, H.L. (2001). Aneurysmal SAH: Cognitive outcome and structural damage after clipping or coiling. Neurology, 56, 1672–1677. Haug, T., Sortegerg, A., Sorteberg, W., Lindegaard, K.L., Lundar, T., & Finset, A. (2007). Cognitive outcome after aneurysmal subarachnoid hemorrhage: Time course of recovery and relationship to clinical, radiological, and management parameters. Neurosurgery, 60(4), 649–657. Haug, T., Sortegerg, A., Sorteberg, W., Lindegaard, K.L., Lundar, T., & Finset, A. (2009). Cognitive functioning and health related quality of life after rupture of an aneurysm on the anterior communicating artery versus middle cerebral artery. British Journal of Neurosurgery, 23(5), 507–515. Henson, R.N., Shallice, T., Josephs, O., & Dolan, R.J. (2002). Functional magnetic resonance imaging of proactive interference during spoken cued recall. Neuroimage, 17, 543–558. Hunt, W.E., & Hess, R.M. (1968). Surgical risk as related to time of intervention in repair of intracranial aneurysm. Journal of Neurosurgery, 28, 14–20. Grote, E., & Hassler, W. (1988). The critical first minutes after subarachnoid hemorrhage. Neurosurgery, 22, 654–656. Kreiter, K.T., Copeland, D.L., Bernardini, G.L., Bates, J.E., Peery, S., Claassen, J., y Mayer, S.A. (2002). Predictors of cognitive dysfunction after subarachnoid hemorrhage. Stroke, 33, 200–209. Larsson, C., Forssell, A., Ro¨nnberg, J., Lindberg, M., Nilsson, L.G., & Fodstad, H. (1994). Subarachnoid blood on CT and memory dysfunctions in aneurysmal subarachnoid hemorrhage. Acta Neurologica Scandinavica, 90(5), 331–336. Larsson, C., Ro¨nnberg, J., Forssell, A., Nilsson, L.G., Lindberg, M., & ¨ ngquist, K.A. (1989). Verbal memory function after subarachnoid A haemorrhage determined by the localisation of the ruptured aneurysm. British Journal of Neurosurgery, 3(5), 549–560. Levine, B., Robertson, I.H., Clare, L., Carter, G., Hong, J., Wilson, B.A., y Stuss, D. (2000). Rehabilitation of executive functioning: An experimental-clinical validation of goal management training. Journal of the International Neuropsychological Society, 6, 299–312. Manning, L., Pierot, L., & Dufour, A. (2005). Anterior and non-anterior ruptured aneurysms: Memory and frontal lobe function performance following coiling. European Journal of Neurology, 12, 466–474. Mavaddat, N., Kirkpatrick, P.J., Rogers, R.D., & Sahakian, B.J. (2000). Deficits in decision-making in patients with aneurysms of the anterior communicating artery. Brain, 123, 2109–2117.
S. Sheldon et al. Mayer, S.A., Kreiter, K.T., Copeland, D., Bernardini, G.L., Bates, J.W., Perry, S., y Connolly, E.S., Jr. (2002). Global and domainspecific cognitive impairment and outcome after subarachnoid haemorrhage. Neurology, 59(11), 1750–1758. Ogden, J., Mee, E.W., & Henning, M. (1993). A prospective study of impairment of cognition and memory and recovery after subarachnoid hemorrhage. Neurosurgery, 33, 572–587. Ørbo, M., Waterloo, K., Egge, A., Isaksen, J., Ingebrigtsen, T., & Romner, B. (2008). Predictors for cognitive impairment one year after surgery for aneurysmal subarachnoid hemorrhage. Journal of Neurology, 255(11), 1770–1776. Østbye, T., Levy, A.R., & Mayo, N.E. (1997). Hospitalization and case-fatality rates for subarachnoid hemorrhage in Canada from 1982 through 1991. Stroke, 28, 793–798. Parkin, A.J., Yeomans, J., & Bindschaedler, C. (1994). Further characterization of the executive memory impairment following frontal lobe lesions. Brain and Cognition, 26, 23–42. Powell, J., Kitchen, N., Heslin, J., & Greenwood, R. (2004). Psychosocial outcomes at 18 months after good neurological recovery from aneurysmal subarachnoid haemorrhage. Journal of Neurology, Neurosurgery, & Psychiatry, 75, 1119–1124. Shimamura, A.P. (1994). Memory and frontal lobe function. In M.S. Gazzaniga (Ed.), The cognitive neurosciences (pp. 803–814). Cambridge: MIT Press. Simard, S., Rouleau, I., Brosseau, J., Laframboise, M., & Bojanowsky, M. (2003). Impact of executive dysfunctions on episodic memory abilities in patients with ruptured aneurysm of the anterior communicating artery. Brain and Cognition, 53(2), 354–358. Strauss, E., Sherman, E.M.S., & Spreen, O. (2006). A compendium of neuropsychological tests: Administration, norms, and commentary. Cambridge, MA: Oxford University Press. Tariq, A., Ai, J., Chen, G., Sabri, M., Jeon, H., Shang, X., & Macdonald, R.L. (2010). Loss of long-term potentiation in the hippocampus after experimental subarachnoid hemorrhage in rats. Neuroscience, 165(2), 418–426. Tidswell, P., Dias, P.S., Sagar, H.J., Mayes, A.R., & Battersby, R.D. (1995). Cognitive outcome after aneurysm rupture: Relationship to aneurysm site and preoperative complications. Neurology, 45, 875–882. Toomela, A., Pulver, A., Romberg, T., Orasson, A., Tikk, A., & Asser, T. (2004). Possible interpretation of subjective complaints in patients with spontaneous subarachnoid haemorrhage. Journal of Rehabilitation Medicine, 36(2), 63–69. Turner, M., Cipolotti, L., Yousry, T., & Shallice, T. (2007). Qualitatively different memory impairments across frontal lobe subgroups. Neuropsychologia, 45, 1540–1552. van Gijn, J., & Rinkel, G.J. (2001). Subarachnoid hemorrhage: Diagnosis, causes and management. Brain, 124, 249–278. Van Overschelde, J.P., Rawson, K.A., & Dunlosky, J. (2004). Category norms: An updated and expanded version of the Battig and Montague (1969) norms. Journal of Memory and Language, 50, 289–335. Wright, A.R., Boeve, B.F., & Malec, J.F. (1999). Amnesia after basal forebrain damage due to anterior communicating artery aneurysm rupture. Journal of Clinical Neuroscience, 6(6), 511–515.