Correspondence : Seung-Hwan Lee, MD, PhD, Department of Psychiatry, Inje University College of Medicine, Ilsan Paik Hospital, 2240 Daehwa-dong, Ilsan-seo-gu, Goyang 411-706, Korea
Tel : +82-31-910-7262, Fax : +82-31-910-7268, E-mail : lshpss@ilsanpaik.ac.kr, lshpss@hanmail.net

Introduction

Alzheimer's disease (AD) represents the most frequent cause of senile dementia. This disease has a slow onset and gradual progression, so its diagnosis is very important. Over the past few years, considerable importance has been attributed to the diagnostic techniques of brain imaging, which are able to provide morphological and functional images. EEG mapping is one of the most widely used methods. In this study, we used qEEG and LORETA^1,2,3 to perform EEG mapping.
qEEG can compare the statistical values of the voltage and frequency that qualitative EEG cannot. Therefore, it can draw a topographic map representing focal brain function. LORETA can make a three dimensional and functional brain map by calculating the brain waves of the scalp surface. Although this method has a low resolution power, it can give us important information about brain functioning.
The EEG patterns of AD patients have consistency. They are characterized by a slowed mean frequency and reduced asymmetry between the two hemispheres.⁴ qEEG has also shown that there is a decrease in the mean frequency along with an increase in the delta and theta power and a parallel decrease in the alpha and beta power in AD patients compared with the corresponding results for normal elderly subjects.^{5,6,7,8,9,10,11} It is generally thought that the earliest changes are an increase in the theta activity and a decrease in the beta activity, which are followed later by a decrease in the alpha activity. The delta activity increases at a later period of the disease course. Patients with severe dementia exhibit a decrease in alpha and an increase in delta activity.^{12,13,14,15,16} 
There have been many equivocal reports about the topographic findings in AD patients. However, Go et al.¹⁷ reported there was a pathophysiologic location especially in the left parietotemporal areas in AD patients compared to the NC group. Moreover, they reported that these findings were comparable with the PET or SPECT findings.^18,19 Duffy et al.²⁰ reported that the areas of maximal group differences in slow waves between the senile AD patient group and their controls involved the midfrontal and anterior frontal lobes, bilaterally. Elmstahl et al.²¹ reported that the delta wave activity was most marked over the posterior regions of the brain in AD patients. Prichep et al.²² reported that there was no localized or lateralized findings, but only diffuse increased theta waves over all brain regions. Schreiter-Gasser et al.²³ reported that there were increased slow waves in the total brain area, but there were localized decreased fast waves in the left parietotemporal area. 
Generally, the asymmetry of the cerebrum is a rather natural phenomenon. However, the asymmetry of the EEG pattern of AD patients is more significant than that of NC subjects.¹⁷ Celsis et al.²⁴ reported that there were lateral asymmetries of the cognitive functions, SPECT and EEG findings in AD patients, but not in controls. Montplaisir et al.²⁵ reported that the degrees of interhemispheric asymmetry calculated by both qEEG and single photon emission computerized tomography (SPECT) were concordant for the parieto-occipital region. Breslau et al.²⁶ reported that AD patients were characterized by a marked delta asymmetry in the temporal regions, which was not seen in the NC groups. 
Thus, the measurement of qEEG or asymmetry of brain waves may be a useful device for the early detection of AD. Therefore, in this study, we attempted to find the specific EEG pattern, hemispheric asymmetry findings and functional brain imaging of AD through qEEG and LORETA.^1,2,3 Also, we analyzed the relations between the mini mental status exam (MMSE) score and the power of each frequency value. 

Methods 

Subjects 
   The AD group consisted of 22 patients (19 female and 3 male) who fulfilled the DSM-IV criteria of dementia of Alzheimer's type. Their mean age was 73.8±7.6 years with a mean duration of AD 22.4±19 months. Patients with other medical conditions known to cause dementia were excluded by means of neurological, serological and imagery tests, including computed tomographic imaging scan (CT-scan) and magnetic resonance imaging (MRI). The symptom severity of AD was assessed by MMSE. The mean MMSE score was 19.2±3.6. The control group consisted of 27 subjects (13 female and 14 male) with no personal history of psychiatric or neurological abnormalities. Their mean age was 66.5±4.7 years and their mean MMSE score was 27.37±1.1 (Table 1). 

EEG Recording and analysis 
   The 18 EEG channels of the international 10-20 criteria were used. The right ear was used as a reference electrode. The measurements were performed with the subjects laying down in a resting position. Their brain waves were recorded about 15 minutes using a Nicolete system (Nicolete biomedical, Madison, WI, WSA) with a sampling rate of 250 Hz/channel, a sensitivity of 7μV, a lower filter of 1 Hz, a higher filter of 70 Hz, and a time constant of 0.3. Five epochs (eye closed state) were taken per subject over the whole record. The length of an epoch was 4.5 seconds, and eye movement and blinking and artifact data were visually screened and rejected. In the analysis of the qEEG, the delta range was 1-3 Hz, the theta range 4-7 Hz, the alpha range 8-12 Hz, and the beta range 13-25 Hz. 

Statistical analysis 
   The independent t-test and bivariate correlation were used to analyze the EEG relative values, including the spectral power and asymmetry. 
To analyze the asymmetry, we used the lateral asymmetry index (LAI).²⁷ The LAI was determined by comparing the corresponding frequency band percentages for the left and right hemispheres. The LAI was computed by dividing the differences between the two hemispheres by their sum, A=(Pleft-Pright)/(Pleft+Pright), where Pleft and Pright are the relative powers of the corresponding frequency band in the appropriate brain region. The resulting values potentially ranged from 1, when the right hemisphere had zero activity, to -1, when the left hemisphere had zero activity. An index of 0 indicated equivalent activity in the two hemispheres. A positive LAI points to dominant brain activity in the left hemisphere, while a negative LAI indicates dominant brain activity in the right hemisphere. 
We made functional source images using the LORETA-key package.^1,2,3 The maximum t-statistics of LORETA is a kind of non-parametric analysis. Thus, it compares each group in 5,000 randomized comparisons. The voxel-byvoxel independent t-test of each group was conducted. We obtained LORETA images that were increased in each frequency band in each brain region. 

Results 

We compared the relative value of the spectral powers in each frequency between the AD and NC groups, and found prominent increases in the theta and delta power spectra in the AD patients compared to the NC subjects. In the theta frequency spectrum, there were significant statistical differences in almost all electrodes in the AD patients compared to the NC subjects. In addition, the AD patients showed maximal differences in the T4, T6, and C4 electrodes (p<0.001) compared to the NC subjects. In the delta frequency spectrum, the AD patients showed statistically significant differences in the T2 electrodes (p<0.01). However, in the alpha and beta frequency bands, there were no statistically significant differences between the AD and NC subjects (Table 2 and 3). 
There was no significant statistical difference in the LAI between the AD and NC groups (Table 4). 
In the LORETA source imaging, increased theta wave activity was observed, especially in Brodmann areas 40, 2, 1, 3 and 5 (inferior parietal lobule, post central gyrus) in the AD patients compared to the NC subjects. Increased delta wave activity was observed especially in Brodmann areas 7, 5, 4, 3 and 40 (post central gyrus) in the AD patients compared to the NC subjects (Figure 1 and 2). 
The MMSE score had a significant negative correlation with the theta waves and a positive correlation with the alpha waves in the AD patients. There was a positive correlation between the duration of illness and theta waves in the AD patients (Table 5). 

Discussion 

We found that the AD patients had significantly increased slow wave activities over all of the electrodes compared with the NC subjects. Moreover, there was a significant correlation between the MMSE and slow waves. These findings suggest that neuronal death in AD can cause increased slow wave activity, which is a pathognomonic feature of this disease. 
The pathophysiological origin of the slowing of the EEG in AD can be explained by the cholinergic deficit. AD is thought to be a syndrome of neocortical disconnection, in which profound cognitive losses arise from the disrupted structural and functional integrity of the long cortico-cortical tracts. Senile plaques and neurofibrillary tangles of AD prominently inhibit the signal transmission of long cortico-cortical association fibers. The neurons are markedly deficient in the basal forebrain nuclei, and this deficiency may account for the severe diminution in the levels of choline acetyltransferase and acetylcholine in the neocortex and paleocortex. The brain of AD patients exhibits a significant reduction in the markers of cholinergic transmission. The atrophy of the basal forebrain cholinergic neurons innervating the neocortex and hippocampus is also observed in AD. Since several studies demonstrated that acetylcholine (Ach) and the basal forebrain system maintain desynchronized EEG activity, a loss of cholinergic innervation of the neocortex might play a critical role in the slowing of the EEG of AD patients.⁴ 
Ricceri et al.²⁸ conducted an animal experiment, in which they examined the long-term effects of neonatal lesions of the basal forebrain cholinergic neurons induced by intracerebroventricular injections of an immunotoxin. In the animals lesioned on postnatal day 7 and tested 6 months later, the EEG cortical patterns presented changes in their alpha, beta and delta activities similar to those observed in Alzheimer-like dementia. These findings indicate that neonatal and permanent basal forebrain cholinergic hypofunction is sufficient to induce behavioral and neuropathological abnormalities. 
There was no significant statistical difference in the LAI of the relative spectral power between the AD and NC groups in our study. However, when we used the absolute spectral power, we could find asymmetry of the EEG power in the AD patients. Most of our AD patients were in the early stage of AD. We believe that this is one of the reasons why there was no asymmetry of the relative spectral power in our subjects. 
In LORETA imaging, maximal different areas of slow wave activities were observed generally in the right parietotemporal region in the AD patients compared to the NC subjects. Even though there is some controversy surrounding this issue, this finding is in agreement with the results of a previous study.²⁰ Duffy et al.²⁰ reported that the maximal group differences between presenile patients and NC subjects are detected in the right posterior temporal area. Moreover, the right-sided numerical features derived from the topographic maps proved most useful in differentiating the presenile patients and their age-matched controls. 
There was a significant negative correlation between MMSE and theta waves, and a significant positive correlation between MMSE and alpha waves in the AD patients. Furthermore, there was a positive correlation between the duration of illness and theta waves in the AD patients in our study. Correlational analyses between the MMSE scores, duration of illness and spectral power of qEEG indicate that increased slowing activity was associated with the state of progression of the disease and the severity of AD. These findings imply that EEG can be a useful diagnostic tool for AD. 
There are several limitations in our study. The first is the age difference between the AD patients and NC subjects. The age of the AD patients was relative higher than that of the NC subjects. Even though there were no significant correlations between the age and slow waves activities in our subjects, the age is surely an important factor influencing EEG activity. Secondly we only used the MMSE to evaluate the severity of symptoms of the AD patients. If we had also used other evaluation scales such as the Dementia Rating Scale, Wechsler Intelligence Test, Wechsler Memory Test, more informative functional correlations in AD would have been found. Thirdly, we could not control the effect of psychotropic drugs. There is some controversy about the effect of psychotropic drugs on EEG. For example, long-term treatment with donepezil led to a lesser deterioration of qEEG,²⁹ which implies less slow wave development. On the other hand, the risk of EEG epileptiform discharge was varied widely among specific antipsychotics.³⁰ Also, Tarn et al.³¹ reported no significant differences in alpha, beta or theta activity after 4 weeks treatment of depressed patients treated with fluoxetine or amitriptyline. The above results indicate that the psychotropic drugs, which our subjects might have been taking, do not have any significant effects on the changes in the slow wave activity of EEG. However, in order to prove this beyond dispute, the effects of psychotropic drugs need to be controlled in future studies. Finally, we utilized 18 channels of EEG for the LORETA functional mapping. To obtain more precise functional brain imaging, a greater number of electrodes needs to be used. 
Even though our study has some limitations, as mentioned above, our results indicated that EEG can be used as a valuable diagnostic tool in AD. We suggest the qEEG and LORETA imaging can be used as an informative tool for the detection of AD. In this cohort, we will try to apply EEG as a possible tool for the early detection of AD in subjects having risk factors. 

1. Pascual-Marqui RD, Esslen M, Kochi K, Lehmann D. Functional imaging with low-resolution brain electromagnetic tomography (LO RETA): a review. Clin Pharmacol 2002;24:91-95. 

2. Babiloni C, Binetti G, Cassetta E, Cerboneschi D, Dal Forno G, Del Percio C, et al. Mapping distributed sources of cortical rhythms in mild Alzheimer's disease. A multicentric EEG study. Neuroimage 2004;22:57-67. 

3. Lee SH, Kwon GH, Park YM, Kim H, Lee KJ, Chung YC. Low resolution electromagnetic tomography (LORETA) source imaging compared with structural brain imaging in patients having organic brain lesion. J Korean Neuropsychiatr Assoc 2006;45:199-207. 

4. Jeong J. EEG dynamics in patients with Alzheimer's disease. Clin Neurophysiol 2004;115:1490-1505. 

5. Bennys K, Rondouin G, Vergnes C, Touchon J. Diagnostic value of quantitative EEG in Alzheimer's disease. Neurophysiol Clin 2001;31:153-160. 

6. Brenner RP, Ulrich RF, Spiker DG, Sclabassi RJ, Reynolds CF 3rd, Marin RS, et al. Computerized EEG spectral analysis in elderly normal, demented and depressed subjects. Electroencephalogr Clin Neurophysiol 1986;64:483-492. 

7. Coben LA, Danziger WL, Berg L. Frequency analysis of the resting awake EEG in mild senile dementia of Alzheimer type. Electroencephalogr Clin Neurophysiol 1983;55:372-380. 

8. Brunovsky M, Matousek M, Edman A, Cervena K, Krajca V. Objective assessment of the degree of dementia by means of EEG. Neuropsychobiology 2003;48:19-26. 

9. Adler G. The EEG as an indicator of cholinergic deficit in Alzheimer's disease. Fortschr Neurol Psychiatr 2000;68:352-356. 

10. Ihl R, Dierks T, Martin EM, Frolich L, Maurer K. Importance of the EEG in early and differential diagnosis of dementia of the Alzheimer type. Fortschr Neurol Psychiatr 1992;60:451-459. 

11. Comi G, Leocani L. Neurophysiological imaging techniques in dementia. Ital J Neurol Sci 1999;20:265-269. 

12. Coben LA, Danziger W, Storandt M. A longitudinal EEG study of mild senile dementia of Alzheimer type: changes at 1 year and at 2.5 years. Electroencephalogr Clin Neurophysiol 1985;61:101-112. 

13. Hier DB, Mangone CA, Ganellen R, Warach JD, Van Egeren R, Perlik SJ, et al. Quantitative measurement of delta activity in Alzheimer's disease. Clin Electroencephalogr 1991;22:178-182. 

14. Rodriguez G, Copello F, Vitali P, Perego G, Nobili F. EEG spectral profile to stage Alzheimer's disease. Clin Neurophysiol 1999;110:1831-1837. 

15. Penttila M, Partanen JV, Soininen H, Riekkinen PJ. Quantitative analysis of occipital EEG in different stages of Alzheimer's disease. Electroencephalogr Clin Neurophysiol 1985;60:1-6. 

16. McKhann G, Drachman D, Folstein M, Katzman R, Price D, Stadlan EM. Clinical diagnosis of Alzheimer's disease: report of the NINCDSADRDA Work Group under the auspices of Department of Health and Human Services Task Force on Alzheimer's Disease. Neurology 1984;34:939-944. 

17. Go HJ, Kim HR, Kim DJ. Spatio-temporal pattern analysis in EEG of Alzheimer's dementia. J Korean Neuropsychiatry Assoc 2000;39:403-411. 

18. Szelies B, Ground M, Herholz K, Kessler J, Wullen T, Heiss WD. Quantitative EEG mapping and PET in Alzheimer's disease. J Neurol sci 1992;110:46-56.

19. Kwa VL, Weinstein HC, Posthumus-Mcyjes EF, van Royen EA, Bour LJ, Verhoeff PN, et al. Spectral analysis of the EEG and 99m-Tc-HMPAO SPECT-Scan in Alzheimer's disease. Biol psychiatry 1993;33:100-107. 

20. Duffy FH, Albert MS, McAnulty G. Brain electrical activity in patients with presenile and senile dementia of the Alzheimer type. Ann Neurol 1984;16:439-448. 

21. Elmstahl S, Rosen L, Gulberg B. Qualitative EEG in elderly pati-ents with Alzheimer's disease and healthy controls. Dementia 1994;5:119-124. 

22. Prichep LS, John ER, Ferris SR, Reisberg B, Almas M, Alper K, et al. Quantitative EEG correlates of cognitive deterioration in the elderly. Neurobiol Aging 1994;15:85-90. 

23. Schreiter-Gasser U, Gasser T, Ziegler P. Quantitative EEG analysis in early onset Alzheimer's disease. Electroencephalogr Clin Neurophysiol 1993;86:15-22. 

24. Celsis P, Agniel A, Puel M, Le Tinnier A, Viallard G, Demonet JF, et al. Lateral asymmetries in primary degenerative dementia of the Alzheimer type: a correlative study of cognitive, haemodynamic and EEG data, in relation with severity, age of onset and sex. Cortex 1990;26:585-596. 

25. Montplaisir J, Petit D, McNamara D, Gauthier S. Comparisons between SPECT and quantitative EEG measures of cortical impairment in mild to moderate Alzheimer's disease. Eur Neurol 1996;36:197-200. 

26. Breslau J, Starr A, Sicotte N, Higa J, Buchsbaum MS. Topographic EEG changes with normal aging and SDAT. Electroencephalogr Clin Neurophysiol 1989;72:281-289. 

27. Jin SH, Na SH, Kim SY, Ham BJ, Lee DH, Lee JH, et al. Hemispheric laterality and dimensional complexity in schizophrenia under sound and light stimulation. Int J Psychophysiol 2003;49:1-15. 

28. Ricceri L, Minghetti L, Moles A, Popoli P, Confaloni A, De Simone R, et al. Cognitive and neurological deficits induced by early and prolonged basal forebrain cholinergic hypofunction in rats. Exp Neurol 2004;189:162-172. 

29. Rodriguez G, Vitali P, De Leo C, De Carli F, Girtler N, Nobili F. Quantitative EEG changes in Alzheimer patients during long-term donepezil therapy. Neuropsychobiology 2002;46:49-56. 

30. Centorrino F, Price BH, Tuttle M, Bahk WM, Hennen J, Albert MJ, et al. EEG abnormalities during treatment with typical and atypical antipsychotics. Am J Psychiatry 2002;159:109-115. 

31. Tarn M, Edwards JG, Sedgwick EM. Fluoxetine, amitriptyline and the electroencephalogram. J Affect Disord 1993;29:7-10.