Diagnostic Value of Periodic Complexes in Creutzfeldt–Jakob Disease Bernhard J. Steinhoff, MD,1,2 Inga Zerr, MD,2 Maya Glatting, MD,2 Walter Schulz-Schaeffer, MD,3 Sigrid Poser, MD,2 and Hans A. Kretzschmar, MD3,4 In 1996, our group published objective electroencephalogram (EEG) criteria to define periodic sharp-wave complexes (PSWCs) suggestive for Creutzfeldt–Jakob disease (CJD). These criteria have since then been strictly applied in all cases reported to us as possible CJD in the course of the German CJD surveillance study. Furthermore, EEG analysis of the records was performed without any additional information on complementary clinical and laboratory data. In this study, we investigated sensitivity, specificity, and the predictive values of these EEG criteria exclusively in cases in which autopsy confirmed (n ⴝ 150) or excluded (n ⴝ 56) CJD. EEG criteria were positive in 64% (n ⴝ 96) of the CJD cases and falsely positive in 9% (n ⴝ 5) of other dementias. The resulting figures for sensitivity, specificity, and positive and negative predictive values were 64%, 91%, 95%, and 49%, respectively. In the falsely positive cases, Alzheimer’s disease (n ⴝ 4) and vascular dementia (n ⴝ 1) were the underlying diseases. However, only in one of these five cases both clinical and EEG data would have led to the false-positive result to diagnose probable CJD. These data prove the high diagnostic value of our objective EEG criteria in CJD. Ann Neurol 2004;56:702–708 Sporadic Creutzfeldt–Jakob disease (sCJD) is the most frequent human prion disease.1–3 Because the diagnosis is proved only post mortem and sCJD always has a fatal course, it is essential to establish diagnostic intra vitam criteria as sensitive and especially as specific as possible. For decades, along with the widely accepted clinical criteria,4 periodic sharp-wave complexes (PSWCs) in the electroencephalogram (EEG) were reported to represent the most typical finding in the course of sCJD5–9 and therefore were included in the World Health Organization classification criteria of sCJD.4 Because PSWCs in sCJD were too poorly defined for objective diagnostic use, especially in epidemiological or multicenter studies, we proposed objective criteria.8 In this previous study, we found an excellent interobserver reliability along with a sensitivity of 67% and a specificity of 86%, respectively.8 The drawback of this study was the relatively small group of patients, with 15 patients in the CJD and 14 in the non-CJD group. In addition, all non-CJD cases were scored according to their clinical course and not confirmed by autopsy. Since then, we have applied our objective EEG criteria as follows. One board-certified electroen- cephalographer (B.J.S.) scored all EEG records of clinically suspected CJD patients reported to the German Surveillance Study Group without any information on clinical and additional neuroradiological or biochemical data. We thus were able to collect a large series of patients, comprising 150 autoptically verified versus 56 autoptically excluded cases. This article addresses the diagnostic value of our objective EEG criteria in this series. From the 1Epilepsy Center Kork, Kehl-Kork; Departments of 2Neurology and 3Neuropathology, Georg-August University of Göttingen; and 4Department of Neuropathology, Ludwig-Maximilians University of Munich, Munich, Germany. Address correspondence to Dr Steinhoff, Epilepsy Center Kork, Landstrasse 1, 77694 Kehl-Kork, Germany. E-mail: email@example.com Patients and Methods EEG records of all patients reported to us as suspected CJD were collected and scored by one of us (B.J.S.) who was not aware of the original EEG report results or any additional diagnostic data of the patients. Based only on the objective EEG criteria previously published8 and shown in Table 1, classification of EEG recordings was as typical or not typical. We analyzed only the data of patients who had undergone autopsy to confirm or rule out CJD. From January of 1996 until August of 2000, it was possible to obtain a definite neuropathological postmortem CJD diagnosis in 330 of 1001 patients reported to us as suspected CJD cases. In 56 cases, CJD was autoptically excluded. Demographic and clinical characteristics of the autopsied CJD patients were not different from the patients not undergoing autopsy. The 56 Received Mar 18, 2004, and in revised form Jul 19. Accepted for publication Jul 19, 2004. Published online Sep 24, 2004, in Wiley InterScience (www.interscience.wiley.com). DOI: 10.1002/ana.20261 702 © 2004 American Neurological Association Published by Wiley-Liss, Inc., through Wiley Subscription Services Table 1. Objective Diagnostic Criteria of Periodic SharpWave Complexes Typical for Sporadic Creutzfeldt–Jakob Disease ● Strictly periodic cerebral potentials, the majority with a duration between 100 and 600 milliseconds and an intercomplex interval between 500 and 2,000 milliseconds ● Generalized and lateralized complexes accepted ● At least five repetitive intervals with a duration difference of less than 500 milliseconds to rule out semiperiodic activity From Steinhoff and colleagues.8 non-CJD cases were completely included in our study, whereas a sample of 150 cases among the neuropathologically verified CJD cases appeared to be sufficient for the intention of this study. These 150 patients did not differ in any aspect from the remaining patients and were randomly selected from our charts. In addition to the EEG and neuropathological data, we considered the clinical data (onset and duration of disease, age at death, and clinical symptoms at onset and during the course of the disease). Sensitivity, specificity, and the (positive and negative) predictive values of the intra vitam EEG diagnosis “typical” or “not typical” of the clinical data and of both criteria were assessed. Statistical tests comprised the 2 test or the Fisher exact tests on a 0.05 level, if applicable. Results Clinical Data Table 2 shows the clinical data of the CJD and the non-CJD group. The mean duration of the disease and the 6- and 12-month survivor rate was considerably higher in the non-CJD group, whereas the mean age did not differ markedly. However, the higher standard deviation for age at onset and mean duration of the disease indicates the less homogenous clinical characteristics in the non-CJD group. Clinical Symptoms in the Autoptically Verified Creutzfeldt–Jakob Disease Cases The clinical symptomatology could be completely verified in 131 of the 150 CJD cases. In the remaining cases, only single clinical symptoms could not be judged sufficiently to be included in the data analysis. The detailed figures are shown in Table 3. The onset of the disease was defined as the time point when first unequivocal neurological, psychiatric, or other clinical symptoms became apparent and were documented. At the beginning of the disease, clinical symptoms were similarly distributed with the exception of visual or oculomotor symptoms which occurred significantly more often in CJD. Conversely, myoclonic jerks appeared significantly less frequently in the CJD group. Distinct clinical differences became more evident in the course of the disease, because, beyond visual or oculomotor symptoms, rapidly progressive dementia, cerebellar symptoms, and akinetic mutism were diagnosed significantly more frequently in CJD cases. The incidence of myoclonic jerks did not differ anymore between the two patient groups. The most frequent neuropathological diagnosis in the non-CJD group was Alzheimer’s disease. The detailed figures are shown in Table 4. Sensitivity, Specificity, and Predictive Values of the Clinical Data Clinical symptoms alone had been sufficient for the correct diagnosis of CJD in 139 cases (93%). In the non-CJD cases, clinical criteria suggested the falsely positive diagnosis of CJD in 34 of 56 cases (61%). Thus, clinical criteria achieved a sensitivity of 93% and a specificity of 39%. The positive predictive value (PPV) was 80% (139/173), and the negative predictive value (NPV) was 67% (22/33). Electroencephalogram Recordings CREUTZFELDT–JAKOB DISEASE CASES (N ⫽ 150). A total of 443 EEG recordings could be assessed. Typical EEG findings were apparent in 96 cases (64%). In this group, a mean of 3.0 ⫾ 1.5 recordings were performed (range, 1–9). The first typical EEG was recorded 3.7 ⫾ 3.1 months after the onset of the disease (0.2–19.2 months); the latest typical EEG was recorded 2.3 ⫾ 3.4 months (range day of death, 17.1 months) before death. Typical EEG findings were always apparent dur- Table 2. Clinical Data Characteristic Female (n) Male (n) Female/male ratio Mean age at onset, range (yr) Mean duration of disease, range (mo) Survivors 6 months after onset of disease (%) Survivors 12 months after onset of disease (%) Total (n) CJD Group Non-CJD Group 92 (61%) 58 (39%) 1.6:1 66 ⫾ 9 (24–85) 7 ⫾ 5.7 (1–25) 45% 15% 150 31 (55%) 25 (45%) 1.2:1 64 ⫾ 16 (26–85) 22 ⫾ 29.4 (1–156) 65% 44% 56 CJD ⫽ Creutzfeldt–Jakob disease. Steinhoff et al: EEG in CJD 703 Table 3. Clinical Symptoms in the CJD and Non-CJD Group Symptom CJD (n ⫽ 150), N (%) Non-CJD (n ⫽ 56), N (%) 87/150 (58) 72/147 (49) 51/146 (35) 10/148 (7) 16/148 (11) 12/148 (8) 0/149 (0) 31/55 (56) 19/53 (36) 9/55 (16) 8/55 (15) 10/53 (19) 11/54 (20) 0/55 (0) n.s. n.s. 0.01 n.s. n.s. 0.015 — 144/146 (99) 124/144 (86) 82/146 (56) 92/147 (63) 104/146 (71) 119/148 (80) 65/147 (44) 40/55 (73) 32/51 (63) 17/54 (31) 30/54 (56) 31/52 (60) 39/55 (71) 15/55 (27) ⬍0.001 ⬍0.001 0.002 n.s. n.s. n.s. 0.03 Onset of disease Rapidly progressive dementia Cerebellar symptoms Visual/oculomotor symptoms Pyramidal signs Extrapyramidal signs Myoclonic jerks Akinetic mutism In the course of the disease Rapidly progressive dementia Cerebellar symptoms Visual/oculomotor symptoms Pyramidal signs Extrapyramidal signs Myoclonic jerks Akinetic mutism p Clinical symptoms at onset (time point when first neurological, psychiatric, or other clinical symptoms became apparent and documented) and during the course of the disease in the CJD and the non-CJD group. Statistical significance was assumed on a 0.05 level, 2-test. The clinical symptomatology could be completely verified in 131 of the 150 CJD cases. CJD ⫽ Creutzfeldt–Jakob disease; n.s. ⫽ not significant. ing one of the first five recordings. Before the recording of PSWCs, the mean EEG number was 0.9 ⫾ 1.2 (range, 0 –5). In 54 CJD cases, typical EEG findings, according to our criteria, were not apparent (36%, mean number of recordings 2.8 ⫾ 1.5; range, 1–9). The first recordings Table 4. Neuropathological Diagnosis in the Non– Creutzfeldt–Jakob Disease Group Diagnosis n % Alzheimer’s disease Neoplasms B-cell lymphoma Glioblastoma Gliomatosis cerebri Paraneoplastic encephalitis Vascular encephalopathies Vascular dementia Multiple cerebral infarctions Inflammatory diseases Cerebral vasculitis Encephalitis Meningitis (⫹ hypoxic cerebral lesion) Progressive encephalomyelitis Neurodegenerative diseases Amyotrophic lateral sclerosis Lewy body dementia Parkinson’s disease Other disease Niemann-Pick’s disease (type C) Congophil angiopathy Leucencephalopathy Toxic alcohol encephalopathy Total 31 7 4 1 1 1 4 3 1 6 2 2 1 1 4 2 1 1 4 1 1 1 1 56 55.4 12.5 704 November 2004 Annals of Neurology Vol 56 No 5 7.1 10.7 7.1 7.1 100 were performed 4.9 ⫾ 4.4 months (range, 0.8 –23.9) after the onset of the disease; the latest ones were performed 2.5 ⫾ 3.0 months (range, 3 days to 15.4 months) before death. In most cases (n ⫽ 52), one to five recordings were performed. The distribution of the number of EEG recordings thus was almost identical as in the CJD group with typical EEG findings. NON–CREUTZFELDT–JAKOB DISEASE CASES (N ⫽ 56). A total of 230 EEG recordings (mean, 4.1 ⫾ 2.9; range, 1–17) could be analyzed. In the non-CJD group without typical PSWCs (n ⫽ 51, 91%), a mean of 4.2 ⫾ 2.9 (range, 1–17) recordings were performed. The first recordings took place 15.4 ⫾ 25.1 months (range, 0.3–141.5) after the onset of the disease, the latest one 3.5 ⫾ 6.6 months (range day of death, 38.9 months) before death. An EEG typical for sCJD was found in five patients (9%); four of them suffered from Alzheimer’s disease. The first clinical signs of dementia appeared in Patient 1 at age 61 years. Three EEGs were recorded, 5, 6, and 9 months after the onset of the disease. The third one showed typical PSWCs. The patient died 3 months later. In Patient 2, two EEGs 4 months after the beginning of the disease had been negative, whereas five further recordings 1 month later contained typical PSWCs. He died another 2 months later. The only EEG in Patient 3 was recorded 1 month after onset and 4 months before death. Figure 1 shows an example of this EEG recording. In Patient 4, autopsy revealed multiple cerebral infarctions. Three EEGs were recorded 28, 24, and 13 days Fig. Electroencephalogram example of Case 3, suffering from neuropathologically verified Alzheimer’s disease. Only one EEG was recorded 4 months before the death of the patient. Note the periodic sharp-wave complexes which had been scored as typical for sporadic Creutzfeldt–Jakob disease cases. before he died. The second recording showed typical PSWCs. Whereas EEG recordings 28, 24, and 23 days before death were negative, a recording 3 days after the first EEG was positive in Patient 5. In only one of the five patients with typical EEG findings, the clinical data also suggested CJD. In the other four cases, because of the atypical clinical data, a misdiagnosis of probable CJD was avoided. Table 5 summarizes the data. Sensitivity, Specificity, and Predictive Values of the Electroencephalogram Data We found, based on the correctly diagnosed CJD cases using only our EEG criteria and the false-positive cases in the non-CJD group, that sensitivity and specificity were 64 and 91%, respectively. The PPV was 95% (96/101), and the NPV was 49% (51/105). Table 5. Non-CJD Patients with EEG Recordings Typical for Sporadic CJD Patient No. Age at Onset (yr) EEG Recordings (n) Typical EEG Recordings (n) Neuropathological Diagnosis 1 2 3 4 5 61 84 76 82 81 3 7 1 3 4 1 5 1 1 1 Alzheimer’s disease Alzheimer’s disease Alzheimer’s disease Multiple cerebral infarctions Alzheimer’s disease CJD ⫽ Creutzfeldt–Jakob disease; EEG ⫽ electroencephalogram. Steinhoff et al: EEG in CJD 705 Sensitivity, Specificity, and Predictive Values of the Combination of Clinical and Electroencephalogram Data Ninety-five of 206 patients were classified as probable (typical clinical findings and typical EEG) and 78 as possible CJD (typical clinical findings but atypical EEG). The remaining 33 patients were excluded because neither the clinical nor the EEG data suggested CJD. In the group of 150 autoptically verified CJD cases, 94 had been classified as probable because of corresponding clinical and EEG data, if both criteria were applied. Sensitivity therefore was 63%. As mentioned earlier, only 1 patient of the 56 autoptically excluded cases fulfilled both the clinical and EEG criteria to be classified as probable CJD, resulting in a specificity of 98%. PPV and NPV were 99% and 49%, respectively. Discussion In a previous study, our EEG criteria were associated with a high sensitivity and specificity and an excellent interobserver reliability documented by a k value of 0.958 that strongly suggested the applicability in multicenter series with various electroencephalographers. Beyond the small groups of patients, the drawback of this investigation was the problem that the control group consisted of patients who had been referred as suspected CJD cases but for whom clinical criteria had been used to rule out CJD. It may have happened that CJD patients with an unusual clinical course were included in this control group, thus causing misleading results. Therefore, in this study we included only cases with neuropathologically verified diagnosis and increased the number of cases considerably. Our data show the applicability of our objective EEG criteria, especially for specificity, which in our opinion is probably more important in the diagnosis of a disease that still has such a devastating prognosis as CJD. Clinical data alone were not sufficient enough to meet this objective. Objective EEG criteria alone were associated with a PPV of 95% and, when added to the clinical criteria, improved the PPV value of the clinical symptoms by remarkable 19% to a value of 99%. PSWCs are not only a sensitive and specific diagnostic indicator in sCJD but also appropriate to differentiate probable sCJD from other prion diseases. With the exception of patients with the codon 200 mutation, they occur only occasionally in familial CJD10 and are not recorded in patients with kuru,11 Gerstmann-Sträussler-Scheinker syndrome,12 fatal familial insomnia,13 and the new variant of CJD.14 Despite the high specificity of our objective EEG criteria, there are still rare cases with rapidly progressive dementia other than sCJD with PSWCs. In our first series8 as well as in this and other studies,15 Alzheimer’s disease has been the most frequent differential 706 Annals of Neurology Vol 56 No 5 November 2004 diagnosis to rule out if PSWCs are recorded. Other infrequent possibilities may be dementia with Lewy bodies15 or vascular dementias as in one patient in this series. It is somewhat satisfying that in this study the combined specificity of the established clinical and our EEG criteria resulted in a remarkable figure of 97%. In the five falsely positive cases based on EEG, only once was probable CJD still diagnosed after consideration of the clinical data. The inclusion of other methods such as the detection of 14-3-3 protein in the cerebrospinal fluid16 –18 and MRI studies should allow an almost safe intra vitam diagnosis of sCJD.19,20 Although still a matter of debate, the 14-3-3 protein has been incorporated in the World Health Organization classification criteria,4 whereas MRI criteria of sCJD have been proposed19 but have not yet been officially recommended.4 MRI reached a sensitivity of 67% and a specificity of 93% in a recently published study.20 It would have been tempting to include it in our analysis. However, although MRI was performed in almost all cases, we did not feel safe enough to include it, because the MRI data were not analyzed by one investigator or at least a homogenous investigator’s group.19 It is interesting to speculate why PSWCs are not found in some sCJD cases. It has been discussed that an imbalance in the subcortical, most probably thalamic, pacemaker systems, may be the underlying condition of PSWCs in CJD.21,22 The crucial role of the ascending reticulothalamocortical system for the generation of PSWCs becomes more evident, if one considers that because of the cortical degeneration in CJD patients the physiological sleep architecture is replaced by a “tracé-alternante”–like pattern. In preterm newborns, this EEG finding of rapidly alternating sleepEEG and PSWCs reflects the still-immature cortical electrophysiological activity.22–24 In adult patients with CJD, the very similar EEG findings result from a dramatic degeneration of the formerly intact cerebral network. PSWCs in CJD may be unilateral.6,24 –28 This may reflect a state of the disease when the commissural progress, which has been shown in experiments,29 has not yet led to the diffuse cortical disease.27,28 Unequivocally, the probability of PSWCs corresponds to the amount of neuronal cell loss.24 The tendency of unilateral PSWC activity to a bilateral spread during the course of CJD has been reported25 and has been seen by us in several instances. The dominating bilateral distribution of PSWCs suggests an underlying midline pacemaker.30 Traub und Pedley31 emphasized the important role of the corpus callosum for widespread EEG activity in CJD and suggested that they originate from a pathological electrophysiological interneuronal coupling based on diffuse neuronal cell destruction during the course of the disease. The short intervals of PSWCs in CJD suggest that the subcortical-cortical electrophysiological axis is still intact but pathologically simplified, whereas underlying cortical deafferentiation typically would result in markedly longer intercomplex intervals.32–34 Consequently, diseases with dominating destruction of the white matter are associated with periodic EEG activity with longer intervals than in CJD.35–38 Finally, the fact that prefinal EEG recordings in sCJD are no longer characterized by PSWCs, but by low-voltage EEG followed by electrocerebral inactivity,39 strongly supports the hypothesis of a still partially functioning subcortical-cortical network as a prerequisite of PSWCs. Moreover, depth electrode EEG of the basal ganglia further support this assumption, because they showed PSWCs in CJD, which were inconstantly apparent on simultaneous surface EEG.26,40 Recent investigations by our group41 support this hypothesis. It could be demonstrated that parvalbuminimmunoreactive (PV⫹) neurons resembling inhibitory neuronal function are markedly reduced in several thalamic regions in sCJD and that in CJD patients with typical PSWCs, a predominant loss of PV⫹ cells was apparent in the reticular thalamic nucleus. In addition, marked PV⫹ cell loss was observed in the brain of patients with homozygosity for methionine at codon 129, namely, in those patients with the highest incidence of PSWCs of 87%.18 Thus, the involvement of the thalamus may be the key to understanding why PSWCs are such a typical finding in sCJD, whereas they are hardly ever found in other human prion diseases. This work was supported by the Deutsche Forschungsgemeinschaft (grant BMG 325-4471-02/15, S.P., H.A.K.). This study was presented in part at the Annual Meeting of the European Neurological Society (ENS), Paris, France, 2001. References 1. Hainfellner JA, Jellinger K, Diringer H, et al. Creutzfeldt-Jakob disease in Austria. J Neurol Neurosurg Psychiatry 1996;61: 139 –142. 2. Will RG, Alpérovitch A, Poser S, et al. Descriptive epidemiology of Creutzfeldt-Jakob disease in six European countries, 1993–1995. EU Collaborative Study Group for CJD. Ann Neurol 1998;43:763–767. 3. Poser S, Mollenhauer B, Krauss A, et al. How to improve the clinical diagnosis of Creutzfeldt-Jakob disease. Brain 1999;122: 2345–2351. 4. World Health Organisation. Consensus on criteria for diagnosis of sporadic CJD. Wkly Epidemiol Rec 1998;73:361–365. 5. Jones DP, Nevin S. Rapidly progressive cerebral degeneration with mental disorder, focal disturbances and myoclonic epilepsy. J Neurol Neurosurg Psychiatry 1954;17:148 –159. 6. Furlan AJ, Henry CE, Sweeney PJ, Mitsumoto H. Focal EEG abnormalities in Heidenhains variant of Jakob-Creutzfeldt disease. Arch Neurol 1981;38:312–314. 7. Levy SR, Chiappa KH, Burke CJ, Young RR. Early evolution of incidence of electroencephalographic abnormalities in Creutzfeldt-Jakob disease. J Clin Neurophysiol 1986;3:1–21 8. Steinhoff BJ, Räcker S, Herrendorf G, et al. Accuracy and reliability of periodic sharp wave complexes (PSWC) in Creutzfeldt-Jakob disease. Arch Neurol 1996;53:162–166. 9. Steinhoff BJ, Riedemann C, Kropp S, et al. Elektroenzephalographische Charakteristika der Creutzfeldt-Jakob Krankheit und ihre Differentialdiagnose. Fortschr Neurol Psychiat 1998;66: 357–365. 10. Brown P, Gibbs CJ, Rodgers-Johnson P, et al. Human Spongiform Encephalopathy: The National Institutes of Health Series of 300 Cases of Experimentally Transmitted Disease. Ann Neurol 1994;35:513–529. 11. Cobb WA, Hornabrook RW, Sanders S. The EEG of kuru. Electroenceph Clin Neurophysiol 1973;34:419 – 427. 12. Prusiner SB. Genetic and infectious prion diseases. Arch Neurol 1993;50:1129 –1153. 13. Zerr I, Giese A, Windl O, et al. Phenotypic variability in fatal familial insomnia (D178N–129M) genotype. Neurology 1998; 51:1398 –1405. 14. Will RG, Zeidler M, Stewart GE, et al. Diagnosis of new variant Creutzfeldt-Jakob disease. Ann Neurol 2000;47:575–582. 15. Tschampa HJ, Neumann M, Zerr I, et al. Patients with Alzheimer’s disease and dementia with Lewy bodies mistaken for Creutzfeldt-Jakob disease. J Neurol Neurosurg Psychiatry 2001; 71:33–39. 16. Hsich G, Kenney K, Gibbs CJ, et al. The 14-3-3 brain protein in cerebrospinal fluid as a marker for transmissible spongiform encephalopathies. N Engl J Med 1996;335:924 –930. 17. Zerr I, Bodemer M, Gefeller O, et al. Detection of 14-3-3 protein in the cerebrospinal fluid supports the diagnosis of Creutzfeldt-Jakob disease. Ann Neurol 1998;43:32– 40. 18. Zerr I, Schulz-Schaeffer W, Giese A, et al. Current clinical diagnosis in Creutzfeldt-Jakob disease: identification of uncommon variants. Ann Neurol 2000;48:323–329. 19. Finkenstaedt M, Szudra A, Zerr I, et al. MR imaging of Creutzfeldt-Jakob disease. Radiology 1996;199:793–798. 20. Schröter A, Zerr I, Henkel K, et al. Magnetic resonance imaging in the clinical diagnosis of Creutzfeldt-Jakob disease. Arch Neurol 2000;57:1751–1757. 21. Nelson JR, Leffman H. The human diffusely projecting system. Arch Neurol 1963;8:544 –556. 22. Bücking PH, Regli F. Die kurze Periodik paroxysmaler Wellenkomplexe im hirnelektrischen und klinischen Verlauf der Creutzfeldt-Jakobschen Erkrankung. Z EEG-EMG 1979;19: 80 – 87. 23. Goto K, Umezaki H, Suetsugu M. Electroencephalographic and clinicopathological studies on Jakob-Creutzfeldt disease. J Neurol Neurosurg Psychiatry 1976;39:931–940. 24. Bortone E, Bertoni L, Giorgi C, et al. Reliability of EEG in the diagnosis of Creutzfeldt-Jakob disease. Electroenceph Clin Neurophysiol 1994;90:323–330. 25. Au WJ, Gabor AJ, Vijayan N, Markand ON. Periodic lateralized epileptiform complexes (PLEDs) in Creutzfeldt-Jakob disease. Neurology 1980;30:611– 617. 26. Chiofalo N, Fuentes A, Gálvez S. Serial EEG findings in 27 cases of Creutzfeldt-Jakob disease. Arch Neurol 1980;37: 143–145. 27. Heye N, Henkes H, Hansen M-L, Gosztonyi G. Focalunilateral accentuation of changes observed in the early stage of Creutzfeldt-Jakob disease. J Neurol Sci 1990;95:105–110. 28. Heye N, Cervós-Navarro J. Focal involvement and lateralization in Creutzfeldt-Jakob disease:correlation of clinical, electroencephalographic and neuropathologic findings. Eur Neurol 1992;32:289 –292. 29. Kim JH, Maunelidis EE. Serial ultrastructural study of experimental Creutzfeldt-Jakob disease in guinea pigs. Acta Neuropathol (Berlin) 1986;69:81–90. Steinhoff et al: EEG in CJD 707 30. Gloor P, Kalabay O, Giard N. The electroencephalogram in diffuse encephalopathies:electroencephalographic correlates of grey and white matter lesions. Brain 1968;91:779 – 802. 31. Traub RD, Pedley TA. Virus-induced electrotonic coupling: hypothesis on the mechanism of periodic EEG discharges in Creutzfeldt-Jakob disease. Ann Neurol 1981;10:405– 410. 32. Ingvar DH. Electrical activity of isolated cortex in the unaesthesized cat with intact brain stem. Acta Physiol Scand 1955; 33:153–167. 33. Creutzfeldt OD, Struck G. Neurophysiologie und Morphologie der chronisch isolierten Cortexinsel. Arch Psychiatr Nervenkr 1962;203:708 –731. 34. Gaches J. Activités périodiques en EEG. Rev EEG Neurophysiol 1971;1:9 –33. 35. Radermecker J, Poser CM. The significance of repetitive paroxysmal electroencephalographic patterns. World Neurol 1960;1: 422– 431. 708 Annals of Neurology Vol 56 No 5 November 2004 36. Cobb W, Hill D. Electroencephalogram in subacute progressive encephalitis. Brain 1950;73:392– 404. 37. Markand ON, Panszi JG. The electroencephalogram in subacute sclerosing panencephalitis. Arch Neurol 1975;32: 719 –726. 38. Tietjen GE, Drury I. Familial Creutzfeldt-Jakob disease without periodic EEG activity. Ann Neurol 1990;28:585–588. 39. Lee RG, Blair RDG. Evolution of EEG and visual evoked response changes in Jakob-Creutzfeldt disease. Electroenceph Clin Neurophysiol 1973;35:133–142 40. Rayport M. Electroencephalographic, corticographic and intracerebral potentials in two anatomically verified cases of Creutzfeldt-Jakob disease. Electroenceph Clin Neurophysiol 1963;15:921. 41. Tschampa HJ, Herms JW, Schulz-Schaeffer WJ, et al. Clinical findings in sporadic Creutzfedlt-Jakob disease correlate with thalamic pathology. Brain 2002;125:2558 –2566.