Abstract
Cognitive defects in migraine have been reported by several authors. These findings however, are controversial. In this study we carried out an investigation on 14 patients with migraine with aura and 16 with migraine without aura according to the International Headache Society criteria. They were submitted to a comprehensive battery of neuropsychological tests. The patients were compared with a control group not significantly different as to age, sex and education. Migraine subjects showed impaired neuropsychological performances only on some cognitive tests. Both groups of patients did worse than the control group on visuo‐spatial memory tasks, while only migraineurs without aura showed significantly impaired verbal memory performances. The memory defects, both on visuo‐spatial and on verbal cognitive tasks, could depend on an impaired recall mechanism. These memory difficulties seem related to strategically and organizationally defective aspects of learning.
Keywords
Introduction
Cognitive dysfunctions in migraine patients have been reported by several authors (1–6). These findings however, are controversial. While some works claimed the presence of deficits (1, 2, 4, 6), others did not confirm these results (3, 5). Zeitlin & Oddy in 1984 found in a group of patients with severe migraine significantly poorer performances in memory and information processing tests. Also, Hooker & Raskin reported neuropsychologic impairment in migraineurs compared with controls. Ardila & Sanchez described 20 patients with classic migraine presenting cognitive defects selected from a group of 200 subjects. Furthermore, Ai in 1992 administered a neuropsychologic battery to patients with classic and common migraine and he found significant differences compared with controls which became worse with the prolongation of disease. However, in 1987 Sinforiani et al. did not confirm the presence of cognitive disturbances in a sample of migraineurs with and without aura. Also, Leijdekkers et al. reported that there was not cognitive impairment in a sample of female migraine patients. A more recent study reported psychomotor dysfunctions in an alternate finger tapping task (7). Also, visual processing could be different in migraineurs with aura due to an over‐sensitivity to visual stimuli (8). These results however, were not confirmed (9). According to Lezak (10), the large variations in the populations of subjects with migraine seem to confound possible real group differences. Other possible confounders may be the different neuropsychological tests chosen and the different selection criteria of experimental samples. Considering these controversial data we carried out further investigation: our study was designed to assess the presence of cognitive defects in a group of patients affected by migraine with and without aura compared with a control group. We were also interested to look for possible differences in neuropsychological performance existing between migraineurs with and without aura. We chose those neuropsychological tests also able, in our opinion, to give qualitative differences in cognitive performance.
Subjects and methods
Subjects
Thirty patients suffering from migraine were studied. Informed consent was obtained. According to the International Headache Society criteria (11), a diagnosis of migraine with aura was made in 14 of them while a diagnosis of migraine without aura was made in 16 others. Exclusion criteria included other types of headache, a history of central or peripheral nervous system disease, trauma, systemic diseases, major psychiatric disorder. Fourteen healthy subjects were selected as a control group. They were acquaintances or relatives of investigators or patients' relatives. They met all of the exclusion criteria and none of the migraine inclusion criteria. The mean age of patients with migraine with aura (11 women and three men) was 32.21 ± 11.40 years and the mean education was 10.71 ± 3.67 years. The mean age and education of subjects with migraine without aura (15 women and one man) were 34.12 ± 12.86 and 10.94 ± 4.23 years, respectively. The mean age and education of the 12 women and two men serving as control group were 33.86 ± 12.45 and 11.21 ± 3.45 years, respectively. No significant differences among the groups were present. Subjects were free of pain and residual effects of a migraine attack when tested. Current prophylactic treatment was recorded.
Methods
The following tests were selected:
Boston scanning test (or Visual Continuous Performance Test). This is a letter cancellation test used to investigate attention; the subject has to mark all the letters A randomly arranged together with other letters (12).
Raven's Progressive Matrices' 47. We used the simplified 36‐item format which requires visual pattern matching and resolution of analogy problems; this test explores conceptual abilities using non‐verbal material (13).
FAS. Also called the Controlled Oral Word Association Test, this consists of three word‐naming trials beginning with F, A and S; it gives a measure of verbal fluency which is associated with frontal lobe functions (14).
Rey Complex Figure. The copy of this widely used picture allows testing of constructional abilities. It is possible to use both quantitative and qualitative scoring systems (15).
Rey Complex Figure Test (recall administration). The immediate and delayed recall of this figure allows measurement of short‐ and long‐term non‐verbal memory. For the three drawings we used the same scoring system ranging from 0 to 36 (15).
Digit span. Used in the Wechsler batteries (intelligence and memory), this comprises two tests (digit forward and digits backward) which involve different abilities. The subject is asked to repeat random number sequences that the examiner reads. Repeating numbers forward seems to be related mostly to attention while the backward task is likely to involve working memory and mental tracking (16).
Corsi Block‐tapping Test. This is a visuo‐spatial test exploring attention and non‐verbal working memory. It consists of nine cubes on a board; the subject is asked to tap the same cubes that the examiner had touched before in an increasing prearranged sequence (17).
California Verbal Learning Test (CVLT). This examines verbal memory and learning strategies together with conceptual abilities. The 16‐word list is presented five times and each time the subject must repeat as many words as possible. After an interference trial there is an immediate recall, a delayed recall (after 20 min) and a final trial of recognition of the 16 words from a longer list of words. The 16 words belong to four categories (fruits, fishes, articles of clothing, tools) and the ability to make clusters (words of the same category recalled together) is also measured. We used the Italian version of the test where fishes took the place of herbs (18).
Hamilton Depression Scale. The first version of this depression rating test was used (19).
Statistical analysis of the scores obtained in neuropsychological tests comparing controls with experimental groups was performed using
Results
The results are reported in Table 1 and Table 2. We found significant differences between controls and both experimental groups only in the short (migraine without aura: F = 4.73, P = 0.04; migraine with aura: F = 5.77, P = 0.02) and long‐term recall of Rey Complex Figure (migraine without aura: F = 7.16, P = 0.01; migraine with aura: F = 5.55, P = 0.02). Moreover, there were significant differences between controls and migraineurs without aura in Boston scanning test (F = 5.03, P = 0.03) and in some subtests of CVLT (short‐term memory, number of clusters at the second repetition and at long‐term recall). The CVLT items were first analysed by
Neuropsychological performances (all tests but California Verbal Learning Test) of controls, and migraineurs with and without aura
STM, Short‐term memory; LTM, long‐term memory.
The P values in parentheses refer to comparisons between patients and controls.
Neuropsychological performances (California Verbal Learning Test) of controls, and migraineurs with and without aura
STM, Short‐term memory; LTM, long‐term memory.
The P values in parentheses refer to comparisons between patients and controls.
Only seven subjects were on prophylactic treatment (flunarizine in three and anti‐depressant drugs in four). Stratified analysis however, excluded possible differences related to prophylactic treatment.
Qualitative differences were shown during the copy of Rey Complex Figure. Most of the controls (92%) used a global strategy (Osterrieth type 1). It consists of beginning by drawing the central rectangle, adding details afterwards. Of migraineurs, 37% exhibited the same pattern; 53% chose Osterrieth type 2, beginning with a detail attached to the central rectangle, or with a subsection of the rectangle and after completing the rectangle and adding details; 10% chose Osterrieth type 3, beginning by drawing the contour of the figure and after adding the internal details. We did not notice any significant differences between the experimental subgroups.
Discussion
Compared with controls, migraine subjects showed impaired neuropsychological performances only on some cognitive tests. Both groups were similarly affected on visuo‐spatial memory tasks, whereas only the migraineurs without aura showed significantly impaired verbal performances and attention defects compared with controls.
Memory dysfunctions in migraineurs have already been reported (1, 2, 4, 6). These cognitive defects however, were not confirmed by other authors (3, 5). Sinforiani et al. (3) did not find significant differences between migraineurs and controls in any of the neuropsychological tasks administered. Also Leijdekkers et al. (5) did not report cognitive impairment in a sample of female migraine patients. The discordances in results between the present and previous studies might be explained by differences in the selection of subjects. As previously suggested (5), it is possible that testing only subjects visiting neurological Departments might be an important selection bias. Another difference between the current study and the former ones is that we used different neuropsychological tests. In a previous work (2) verbal memory impairment in migraineurs was demonstrated only in the most difficult test of recall. Furthermore, visuo‐spatial memory was investigated visually, presenting four geometric figures without asking any copy of the material. Also, Rey 15 words and Wechsler Memory Figures, used by other authors (3), are in some ways different from the tests we used. The memory tests that we administered (CVLT and Rey Complex Figure Test recall administration) appear to be more sensitive in exploring strategical problems. In the learning trials in fact, subjects showed more difficulties both in verbal and non‐verbal tasks. Memory defects, both on visuo‐spatial and on verbal cognitive tests, could depend on an impaired recall mechanism. Good recognition in CVLT confirmed this assumption. These memory difficulties seemed to be related to strategically and organizationally defective aspects of learning: as a matter of fact a large part of both experimental groups used a different and less effective strategy of copy of Rey Complex Figure compared with controls. This concept was also stressed by results obtained in CVLT, which showed a reduced clustering in migraineurs, even if significant differences were found only in migraineurs without aura. A possible explanation could be found in a right hemisphere dysfunction: lack of global processing, visual memory and attention impairment should support this hypothesis. Also the poor retrieval may be explained by right prefrontal dysfunction (20). Finally, we can not exclude that other clinical variables, different from illness duration, might play a significant role in neuropsychological dysfunctions. Lack of a significant correlation between illness duration and cognitive dysfunctions seems to indicate that alterations in processing mechanisms of information might be present also in the early stages of migraine. In conclusion, further studies should include populations sharing similar clinical features who should be submitted to the same neuropsychological tests.