We related the histopathologic changes associated with incidental white matter signal hyperintensities on MRIs from 11 elderly patients (age range, 52 to 82 years) to a descriptive classification for such abnormalities. Punctate, early confluent, and confluent white matter hyperintensities corresponded to increasing severity of ischemic tissue damage, ranging from mild perivascular alterations to large areas with variable loss of fibers, multiple small cavitations, and marked arteriolosclerosis. Microcystic infarcts and patchy rarefaction of myelin were also characteristic for irregular periventricular high signal intensity. Hyperintense periventricular caps and a smooth halo, however, were of nonischemic origin and constituted areas of demyelination associated with subependymal gliosis and discontinuity of the ependymal lining. Based on these findings, our classification appears to reflect both the different etiologies and severities of incidental MRI signal abnormalities, if it is modified to treat irregular periventricular and confluent deep white matter hyperintensities together.

We studied 52 asymptomatic subjects using magnetic resonance imaging, and we compared age-matched groups (51-70 years old) with and without white matter lesions with respect to carotid ultrasonography, cerebral blood flow (xenon-133 injection), and cerebrovascular risk factors. In the group with white matter signal abnormalities, we noted a higher frequency of extracranial carotid artery disease, a lower mean gray matter blood flow (Fl), and a significant reduction (p<0.05) in blood flow of the slow-flowing (F2) compartment. Hypertension, diabetes mellitus, and cardiac diseases (p<0.002) were found more often in this group. Our results indicate that a higher incidence of changes known to be associated with an increased risk for stroke exists in the presence of white matter lesions in normal elderly individuals. 12 as well as in normal aging. 34 The frequency of these hyperintense white matter foci on T2-weighted images increases with advancing age 2 and with the number of cerebrovascular risk factors.5 A correlation between the presence of white matter lesions (WMLs) and cerebrovascular disease has therefore been suggested, '-35 implying that magnetic resonance signal abnormalities in asymptomatic subjects may provide evidence of "silent" cerebrovascular disease. 6To test further the significance of WMLs, we investigated their association with factors known to increase the risk for stroke 7 -10 in a population of symptom-free individuals. To our knowledge, ours is the first study to evaluate normal subjects with respect to cerebrovascular changes using magnetic resonance imaging (MRI), the xenon-133 injection method of cerebral blood flow (CBF) measurement, and carotid ultrasonography simultaneously. Subjects and MethodsThe 52 subjects (mean age 58, range 31-78 years) we used are a sample of volunteers participating in a prospective field study on the incidence of cerebrovascular risk factors in the population of Graz, Austria, and the surrounding region. Subjects were eligible for the study if they had no evidence of cerebrovascular disease on history or neurologic examination and after exclusion of other neurologic or psychiatric disorders. Subjects were free of symptoms of systemic diseases and had not come to medical attention for cerebrovascular risk factors. MRI and extracranial Doppler ultrasonography were performed in all subjects, and CBF was measured in 32 subjects; these tests were carried out within 1 day. Evidence of hypertension (blood pressure of >160 mm Hg systolic), diabetes mellitus (fasting blood sugar concentration of >160 mg/dl), and cardiac diseases (coronary heart disease, cardiomyopathy, arrhythmias, etc.) was based on appropriate findings on two independent visits.All subjects were studied with a superconducting magnet at a field strength of 1.5 T (Gyroscan S15, Philips, Eindhoven, The Netherlands) and the spin-

We reviewed the MRIs of 49 asymptomatic volunteers (age range, 31 to 77 years) and of 50 MS patients (age range, 14 to 63) for areas of increased signal (AIS) and features discriminating MS lesions from lesions seen with normal aging. We obtained optimal specificity of MRI interpretation (100%) if we required at least two of the following three AIS features--size greater than or equal to 6 mm, abutting ventricular bodies, infratentorial location--for a positive MRI diagnosis of MS. Applying these criteria to the MRIs of elderly patients with suspected MS should significantly improve specificity (p less than 0.001) over current quantitative criteria (at least three AIS greater than or equal to 3 mm) without significantly decreasing sensitivity.

To determine the effects of MRI white matter hyperintensities (WMH) on cognitive functioning, we used neuropsychologic tests and MRI to study 150 elderly volunteers free of neuropsychiatric or general disease. There were 76 (50.3%) individuals without and 74 (49.7%) with WMH. The latter subset was older (61.3 +/- 6.6 years versus 58.5 +/- 5.8 years, p = 0.005), had a higher mean arterial blood pressure (103.7 +/- 11.4 mm Hg versus 99.9 +/- 10.3 mm Hg, p = 0.03), and a larger ventricular-to-intracranial-cavity ratio (6.3 +/- 5.6% versus 4.7 +/- 1.6%, p = 0.02). Individuals with WMH performed worse than their counterparts without such abnormalities on all tests administered. After adjusting for the group differences in age, arterial blood pressure, and ventricular size, we noted statistically significant results on form B of the Trail Making Test (121.8 +/- 37.8 msec versus 100.3 +/- 47.9 msec, p = 0.04), a complex reaction time task (680.8 +/- 104.9 msec versus 607.1 +/- 93.9 msec, p = 0.001), and the assembly procedure of the Purdue Peg-board Test (27.5 +/- 5.8 versus 30.6 +/- 5.9, p = 0.02). Partial correlations did not reveal any relationship between test scores and the semiautomatically assessed total area of WMH. Our data suggest that the presence of WMH exerts a subtle effect on neuropsychologic performance of normal elderly individuals, which becomes particularly evident on tasks measuring the speed of more complex mental processing.

Studies on the prevalence of MRI signal abnormalities in the brains of migraineurs have yielded controversial results. In order to provide further data on this issue we reviewed the MRI scans of 38 migraine patients without current neurologic symptoms (mean age 35.8 +/- 11.9 years). In addition, we compared the findings in those 24 migraineurs under 50 years without major cerebrovascular risk factors (mean age 30.1 +/- 9.0 years) to that in 14 headache and risk factor free volunteers (mean age 37.8 +/- 5.3 years). Overall, focal areas of hyperintense signal were seen in 15 (39%) patients. They were present on both proton density and T2-weighted spin-echo sequences. Lesion prevalence varied according to the type of headache (18% in migraine without aura, 53% in migraine with typical aura, 38% in basilar migraine). The subset of migraine patients under 50 years exhibited MRI signal abnormalities more than twice as often as controls (33% vs. 14%). Punctate white matter hyperintensities were the predominant finding and were seen in 10 of 15 individuals with MRI lesions. More striking signal abnormalities consisted of symmetrical areas of hyperintensity lateral to the posterior horns in two 24 year old patients and of extensive white matter damage with lacunar infarcts in a 59 year old woman. Our findings confirm a higher prevalence of MRI lesions in a mixed group of migraineurs than in headache free individuals. Signal abnormalities are most often non-specific, however their occurrence relates to the type of migraine.