Correspondence: Doh Kwan Kim, M.D., Ph.D., Department of Psychiatry, Samsung Medical Center, Sungkyunkwan University School of Medicine, 50 Irwon-dong, Kangnam-gu, Seoul 135-710, Korea
Tel: +82-2-3410-3582, Fax: +82-2-3410-0941, E-mail: paulkim@smc.samsung.co.kr

Introduction

Major depressive disorder (MDD) is one of the most common mental disorders, which is characterized by episodic depression, loss of interest in daily life, and despair. Cho et al. reported that the lifetime prevalence of MDD in the Korean population ranges up to 4.25%.¹ The worldwide lifetime prevalence of MDD ranges from 4.4% to 20%,^2,3 and MDD is predicted to become the second most devastating disease in regard to its social and financial burden by 2020.^3,4,5
The etiology of MDD remains unknown. Since Schildkraut proposed the catecholamine hypothesis for affective disorders, based on the hypothesized deficiency of biological amines at the synapses of neurons in the central nervous system,⁶ a variety of studies have been performed and therapeutics such as selective serotonin reuptake inhibitor (SSRI) and serotonin-norepinephrine reuptake inhibitor (SNRI) have been developed for treatment.^7,8,9 However, the monoamine hypothesis has its limits, as demonstrated by the following clinical observations. With the administration of antidepressants, it takes 10-14 days to show a clinical response, and another 4-6 weeks before a medication trial can be ruled a failure, yet initial drug treatment fails in 30-40% of MDD patients.^3,10,11 Therefore, to further develop the biological monoamine hypothesis beyond the level of the neurotransmitter (NT) or receptors, Duman et al. suggested that the mechanism of action should be studied at the molecular and cellular levels and might thus be found to be associated with the intracellular signal transduction pathway, which was linked to the expression of specific genes.^12,13,14 Accordingly, related studies have shown that cAMP response element-binding protein (CREB) stimulates the expression of specific genes and resulting neurotrophic factors as final functional substances and that mediators of the therapeutic response to antidepressants are involved in neuroplasticity and neuronal survival by stimulating neurogenesis and synaptic plasticity.^15,16,17 Consequently, in regard to the action mechanism of antidepressants and the pathophysiology of depression, attention has been now focused on one of the representative neurotrophic factors: brain-derived neurotrophic factor (BDNF).^{17,18,19,20,21,22,23}
To our knowledge, most of the BDNF in the blood is stored in platelets, and BDNF can freely and bidirectionally cross the blood-brain barrier (BBB) and is normally in equilibrium among three compartments: platelets, plasma, and brain.^24,25,26 Based on these characteristics, most of studies conducted until now on the basis of the neuroplasticity hypothesis have targeted serum BDNF to represent the BDNF stored in platelets and have reported results that are compatible with the neuroplasticity hypothesis.^{27,28,29,30,31}
Considering the physiologic equilibrium among the three compartments and the direct contact and interaction between plasma and brain through the BBB, circulating BDNF in the plasma may play more a active role in reflecting neuroplastic change.
We therefore investigated whether plasma BDNF could play a significant role as a biological marker in reflecting neuroplastic change related to the pathophysiology of depression and the action mechanism of anti-depressants, with a case-control study design and a 6- week administration of fluoxetine. In this study, we assumed the following three hypotheses. First, the baseline plasma BDNF level in MDD patients is lower than that of a control group. Second, the severity of depression at baseline is negatively correlated with the baseline plasma BDNF level in MDD patients. Third, with fluoxetine administration, the change in plasma BDNF is much greater in the responder group and is positively correlated with symptom improvement.

Methods and Materials

Subjects

Patients with MDD

The MDD patients in this study were aged 20 years or over and were recruited among patients who visited the outpatient clinic or had been admitted to the department of psychiatry at Samsung Medical Center from February 2003 to November 2004. All patients fulfilled the criteria of the Diagnostic and Statistical Manual of Mental Disorders (DSM-IV) for MDD.³³ The 17-item Hamilton Rating Scale for Depression (HAM-D-17)^34,35,36 was used to measure the severity of depression and a minimum baseline HAM-D-17 score of 15 was set as criterion for study inclusion to ensure all subjects had MDD severity of at least moderate. For all psychiatric diagnoses, the Korean version of the Structured Clinical Interview for DSM-IV Axis-I disorder (SCID-IV) was used.³⁷ Accordingly, the patients with Axis I psychiatric disorders other than MDD, such as alcohol or drug dependence, and with a clinically significant personality disorder were excluded from the study. For the patients older than 60 years, the Korean version of the Mini-mental state Examination (K-MMSE)^38,39 was performed in order to rule out early dementia and the patients with a score lower than 25 points were excluded. Other exclusion criteria were other neurological diseases, significant medical conditions and abnormal laboratory baseline values. Women taking oral contraceptives, menstruating or in pregnancy were excluded. Patients who had taken antidepressants within 2 weeks prior to participating in the study and who could not tolerate the drug or continue the study for other reasons were also excluded. Fluoxetine monotherapy was administered for all patients at the following dose for 6 weeks. Considering the tolerability of the patients, the fluoxetine dose was started at 10-20mg/day and then increased to 20-40mg/day within 2 weeks. The final daily median dosages were 30mg/day of fluoxetine, which is a typical clinical dose for Asian populations. Other antidepressants or psychotropic drugs were not permitted except Lorazepam 1-2mg at bedtime for insomnia. Written informed consent was obtained from all participants. This study was approved by the Institutional Review Board (IRB) of Samsung Medical Center, Seoul, Korea.

Normal control group

Among volunteers recruited through advertisements, individuals were selected who were age- and sex-matched to the MDD patients and satisfied the criteria of total score on the Beck Depression Inventory (BDI) (<19 points for males and <22 points for females).^40,41 Exclusion criteria were a past history of psychiatric disorder or severe physical diseases such as nervous system diseases, renal disease, hepatic disease, hematological diseases, and thyroid disease. The Minnesota Multiphasic Personality Inventory (MMPI)^42,43 was performed and those with a clinical MMPI index outside the normal range of 35-65 points were also excluded. All participants in the control group provided informed consent.

Study methods

Evaluation of depression severity and responsiveness

The severity of depression symptoms before the administration of the antidepressant, fluoxetine, and after 6-week treatment was measured by an independent clinical psychologist using HAM-D-17. To evaluate the responsiveness to the antidepressant, a 'responder' was defined as one who achieved at least a 50% reduction in baseline HAM-D-17 score after 6-week treatment. The criteria for the 'remitter' was a final score of 7 points or lower.¹⁰ In other words, treatment outcome was measured as response (≥50% improvement on HAM-D-17) and remission (final HAM-D-17 score ≤7).

Measurement of plasma BDNF

Ten-milliliter venous blood samples were collected in an EDTA-vacutainer from MDD patients and normal control volunteers at 9:00~12:00 am. Plasma was collected after serial centrifugation at 3,000 rpm for 15min and 10,000g for 10min at 4 C, and then stored at -70 C until use. Plasma BDNF levels were detected using the BDNF Emax ImmunoAssay system (Promega, Madison, WI, USA). In brief, 96-well plates were coated with anti-BDNF monoclonal antibody after overnight incubation at 4 C. The next day, the plates were washed with wash buffer and blocked at room temperature (RT) for 1 hour without shaking. After washing once, BDNF serial standards and samples were added to the wells and incubated for 2 hours at RT with shaking. After washing 5 times, the captured BDNFs were bound with the second specific BDNF polyclonal antibody for 2 hours at RT with shaking. After washing 5 times, they were detected with species-specific anti-IgY antibody conjugated to horseradish peroxidase, were incubated with chromogenic substrate, and then the reaction was stopped with 1N HCl. The absorbencies were measured at 450nm using an automatic ELISA microreader. Inter-assay and intra-assay variations were less than 5 and 8%, respectively. Additionally, the average values of plasma BDNF measurements performed in triplicate on the same day were used for the study. The standardization curve in each plate was adopted for the measurement.

Statistical analysis
The between-group differences in demographic data and plasma BDNF levels at baseline were analyzed statistically with the Chi-square test or student t-test and Wilcoxon's rank sum test. For correlation between HAM-D-17 scores and plasma BDNF levels at baseline, Spearman's correlation test was used. After treatment, the significance of the change in the plasma BDNF for the MDD group was analyzed with the Wilcoxon's signed rank test, and the difference in the change of plasma BDNF was compared between responders and non-responders using the student t-test or Mann-Whitney test. Significance was set at p<.05.

Results

Demographic and clinical characteristics of MDD patients and normal controls
Forty-nine MDD subjects and 34 normal controls completed the study and were analyzed. The demographic and clinical characteristics of all subjects are presented in Table 1. The mean age was 65.2±10.33 years for the MDD group and 64.6±4.82 years for the normal control group, with no significant difference between the two (t=-0.33, p >0.05). The percentage of female subjects in each group was 73.5% and 55.9%, respectively, but there was no statistically significant, between-group difference (X 2 =2.78, df=1, p=0.095). For the baseline plasma BDNF, the mean level of 1,061 ± 663.70 pg/ml in the MDD group was lower than 1,095±595.64 pg/ml in the normal control group, but this difference was not statistically significant (Z=0.54, p=0.588).

Assignment of responsiveness and the clinical characteristics
After 6-week treatment, 49 MDD patients were evaluated and assigned to the responder (28 patients, 57.1%) and nonresponder (21 patients, 42.9%) according to the pre-determined criteria of responsiveness. The mean age was 64.57±10.72 years for responders and 66.04±10.00 years for nonresponders (Z=-1.05, p=0.293). There were no significant differences between the two groups in gender, age of onset, number of episodes, duration of illness and number of the patients with family history of MDD. Furthermore, there were no significant, between-group differences in weight at baseline and 6-week. As Table 1 illustrates, there was no significant between-group difference in baseline HAM-D score, with mean scores of 20.39± 3.52 and 20.04±2.90, respectively (t = 0.37, p = 0.717). The mean BDNF level at baseline for the responders, 951.37±588.37 pg/ml, was lower than that for the non-responders, 1208.45±741.75 pg/ml, but the difference was not significant (Z= -1.44, p =0.149). Similarly, the difference of 6-week BDNF level between the two groups was not statistically significant (Z= -1.14, p= 0.254). As expected, the mean score HAM-D-17 score at 6-weeks (6.14 vs. 14.00 for responders and nonresponders, respectively) showed a significant difference (Z=- 5.25, p<0.01).

Correlation analysis
In all MDD patients, the correlation between the baseline HAM-D-17 score, reflecting the severity of depression, and the baseline plasma BDNF level was analyzed. On the scatter plot (data not shown), most of the patient cases were widely distributed in the lower part of the plasma BDNF axis, showing a weakly posi-tive correlation. However, Spearman's correlation analysis showed no statistically significant relationship (coefficient =0.213, p=0.142).
Furthermore, there was no statistically significant correlation between the number of episodes and base-line plasma BDNF (coefficient = 0.098, p = 0.499).

The change of plasma BDNF before and after fluoxetine treatment in MDD
After fluoxetine administration, the total change of plasma BDNF in all MDD patients showed a slight decrease without statistical significance (Wilcoxon's signed rank test, S=-186.5, p =0.063). However, the increase of weight was observed with no statistically significant difference (paired student t-test, t =1.781, p=0.081). Although the weight change was considered as a confounding factor, the difference of plasma BDNF between before and after treatment was not significant. Figure 1 presents the change of plasma BDNF in the MDD group, comparing the distribution of plasma BDNF at baseline and 6 weeks. Furthermore, the degree of change of the plasma BDNF was negatively correlated with the age of onset and baseline plasma BDNF (Spearman correla-tion coefficient=-0.378, p<0.01; -0.425, p<0.05, respectively), and positively correlated with the base-line HAM-D score (coefficient=0.281, p<0.05).
Comparing the difference in the change of plasma BDNF according to the responsiveness after 6-week antidepressant treatment, there was no significant difference between responders and nonresponders (t=-0.116, p=0.908). The difference in the change of weight between the two groups was analyzed as a possible confounding factor, but the difference remained statistically not significant (student t-test, t=1.293, p =0.202). Because there was also no significant relationship between the change of weight and the change of plasma BDNF, adjustment of the weight change as a confounding factor was not considered. Moreover, in the absence of any other confounding factors related to both the change of plasma BDNF and responsiveness, no further correction was considered.

Discussion

In this study, we investigated whether plasma BDNF can function as a peripheral biological marker with characteristics different from serum BDNF, and whether these characteristics allow it to more accurately reflect the change of brain neuroplasticity after 6-week administration of the antidepressant, fluoxetine.
Based on the neuroplasticity theory related to the pathophysiology of depression and the action mechanism of antidepressants, many researchers have reported that brain neuroplasticity is both significantly and consistently reflected by serum BDNF.^{27,28,29,30,31} However, to our knowledge, the present study is the first to investigate the potential role and characteristics of plasma BDNF as an intermediate compartment between platelets and the brain in reflecting the neuroplasticity with a 6-week case-control study design.
The findings in this study were not compatible with the three hypotheses proposed in this study based on the neuroplasticity theory and the results reported with serum BDNF in previous studies.^27,30,31 Firstly, the plasma BDNF levels in the MDD patients were not different from those of normal controls, which does not support the findings reported by Karege et al..³² Secondly, the severity of depression was not significantly correlated to the plasma BDNF levels. According to the neuroplasticity hypothesis, the failure of neuroplasticity in the brain is causally related to the pathophysiology of depression. Therefore, in MDD patients, the plasma BDNF levels would be expected to be lower than those in normal controls and to be negatively correlated with the clinical severity of depression as measured by the HAM-D-17 scale, if the plasma is the appropriate compartment which accurately reflects the neuroplastic changes of brain. Considering there were no possible confounding factors such as age and gender which could have affected the circulating BDNF levels in plasma,²⁵ the findings in this study suggest that plasma BDNF itself is not a significant peripheral biological marker that reflects neuroplasticity related to the pathophysiology of depression. However, the effect of weight was not treated as a confounding factor in the comparison between MDD patients and normal controls, which necessitates caution in the final interpretation.
For plasma BDNF to be a predictive factor for responsiveness and to be a biological marker that reflects neuroplasticity related to the action mechanism of antidepressant treatment, plasma BDNF would need to increase in the drug-treated MDD group, and this increase would have to be much more significant in the group showing response after anti-depressant treatment than in the group with little or no response. However, the findings in this study did not support this hypothesis. In contrast, there was a rather modest trend for plasma BDNF to decrease in MDD patients after the 6-week administration of fluoxetine. Considering the absence of any significant increase of circulating plasma BDNF level in the responder group, this suggests that the plasma compartment, different from the platelets, does not reflect or is not so sensitive to the changes of BDNF related to brain neuroplasticity.
The limitations of the study results in not supporting any of the three proposed hypotheses based on the neuroplasticity theory may have been caused by the following factors. First, the quantity of circulating BDNF in the plasma compartment was nearly 100-fold lower than that in the platelet compartment.^32,44 Therefore, the sensitivity of plasma BDNF for detecting the change in BDNF related to neuroplasticity may have been reduced. Second, the sources for the synthesis and secretion of circulating BDNF are diverse, including vascular endothelial cells, smooth muscles, neurons and glia within the central nervous system; therefore, the source is not confined to only neurons in the brain.²⁴ Accordingly, regardless of the neuroplastic change related to the responsiveness to fluoxetine administration, other sources of BDNF might affect the circulating BDNF levels by masking the true effects of the antidepressant on responsiveness and neuroplasticity. Third, confounding factors, either specified or unspecified, might have affected the circulating plasma BDNF levels.²⁵ In this study, two specified confounding factors, age and gender, were controlled, but other specified factors, such as menstrual cycle and weight, and other unspecified confounding factors were only partially or insufficiently considered. Particularly, differences in the exposure to stress,⁴⁵ physical activity⁴⁶ or food⁴⁷ have been reported to influence the neuronal BDNF synthesis and secretion, and subsequently affect the variance of plasma BDNF levels. Finally, it may be necessary to consider in the final interpretation that plasma BDNF in this study was rather lowly distributed (range, 242pg/ml-2543pg/ml), in divergence from that previously reported (range, 1200pg/ml-4700pg/ml).³²
However, this study had several important strengths. First, the sample size was relatively large compared to previous studies.^27,28,29,30 Second, unlike previous studies, one specific antidepressant, fluoxetine, and only plasma BDNF, not serum BDNF, were adopted for the study, as proposed by Gervasoni et al..²⁹ Furthermore, this study design was the first to measure the change of BDNF in the plasma compartment after 6-week antidepressant treatment, in contrast to the previous study conducted by Karege et al.³²
In conclusion, the findings in this study do not suggest that the circulating BDNF level in the plasma compartment plays a significant role as a potential candidate bio-marker to reflect neuroplastic change related to the pathophysiology of depression and the action mechanism of antidepressant treatment, at least with the antidepressant, fluoxetine. Furthermore, it is inferred that the plasma compartment may not directly reflect brain activity, as the platelet compartment does, represented by serum BDNF. However, comprehensive consideration of the results reported by previous studies and the limitations of the present study suggest that the possibility of false negative findings cannot be excluded. Therefore, further studies, in which the limitations in the present study are diminished or resolved, are required to replicate the present findings. In addition, a study design with serial measurement of plasma BDNF would be advisable.

1. Cho MJ, Hahm BJ, Kim JK, Park KK, Chung EK, Suh TW, Kim SU, Cho SJ, Lee JY, Hong JP, Choi YS, Park JI, Lee DW, Lee GC, Bae JN, Shin JH, Chung IW, Park JH, Bae A, Lee CK. Korean Epidemiologic Catchment Area (KECA) Study for Psychiatric Disorders: Prevalence of Specific Psychiatric Disorders. J Korean Neuropsychiatr Assoc 2004; 43 (4): 470-480.

2. Angst J. Epidemiology of depression. Psychopharmacology (Berl) 1992; 106 (suppl): 71-74.

3. Bakish D. New standard of depression treatment: remission and full recovery. J Clin Psychiatry 2001; 62 (suppl 26): 5-9.

4. Murray CJ, Lopez AD. Global mortality, disability, and the contribution of risk factors: Global Burden of Disease Study. Lancet 1997; 349 (9063): 1436-1442.

5. Murray CJ, Lopez AD. Alternative projections of mortality and disability by cause 1990-2020: Global Burden of Disease Study. Lancet 1997; 349 (9063): 1498-1504.

6. Schildkraut JJ. The catecholamine hypothesis of affective disorders: a review of supporting evidence. Am J Psychiatry 1965; 122 (5): 509-522.

7. Meltzer H. Serotonergic dysfunction in depression. Br J Psychiatry 1989; 8 (suppl): 25-31.

8. Dinan TG. Noradrenergic and serotonergic abnormalities in depression: stress-induced dysfunction? J Clin Psychiatry 1996; 57 (suppl 4): 14-18.

9. Hirschfeld RM. History and evolution of the monoamine hypothesis of depression. J Clin Psychiatry 2000; 61 (suppl 6): 4-6.

10. Nierenberg AA, Amsterdam JD. Treatment-resistant depression: definition and treatment approaches. J Clin Psychiatry 1990; 51 (Suppl): 39-47.

11. Hindmarch I. Expanding the horizons of depression: beyond the monoamine hypothesis. Hum Psychopharmacol 2001; 16 (3): 203-218.

12. Duman RS, Heninger GR, Nestler EJ. A molecular and cel-lular theory of depression. Arch Gen Psychiatry 1997; 54 (7): 597-606.

13. Duman RS, Malberg J, Nakagawa S, D'Sa C. Neuronal plasticity and survival in mood disorders. Biol Psychiatry 2000; 48 (8): 732-739.

14. Sandler M. Neurotrophins: possible role in affective disorders. Hum Psychopharmacol 2001; 16 (3): 61-64.

15. Nibuya M, Nestler EJ, Duman RS. Chronic antidepressant administration increases the expression of cAMP response element binding protein (CREB) in rat hippocampus. J Neurosci 1996; 16 (7): 2365-2372.

16. Chen AC, Shirayama Y, Shin KH, Neve RL, Duman RS. Expression of the cAMP response element binding protein (CREB) in hippocampus produces an antidepressant effect. Biol Psychiatry 2001; 49 (9): 753-762.

17. D'Sa C, Duman RS. Antidepressants and neuroplasticity. Bipolar Disord 2002; 4 (3): 183-194.

18. Chen B, Dowlatshahi D, MacQueen GM, Wang JF, Young LT. Increased hippocampal BDNF immunoreactivity in subjects treated with antidepressant medication. Biol Psychiatry 2001; 50 (4): 260-265.

19. Kempermann G, Kronenberg G. Depressed new neurons-- adult hippocampal neurogenesis and a cellular plasticity hypothesis of major depression. Biol Psychiatry 2003; 54 (5): 499-503.

20. Malberg JE, Duman RS. Cell proliferation in adult hippocampus is decreased by inescapable stress: reversal by fluoxetine treatment. Neuropsychopharmacology 2003; 28 (9): 1562-1571.

21. Duman RS. Role of neurotrophic factors in the etiology and treatment of mood disorders. Neuromolecular Med 2004; 5 (1): 11-25.

22. Hashimoto K, Shimizu E, Iyo M. Critical role of brain-derived neurotrophic factor in mood disorders. Brain Res Brain Res Rev 2004; 45 (2): 104-114.

23. Altar CA. Neurotrophins and depression. Trends Pharmacol Sci 1999; 20 (2): 59-61.

24. Fujimura H, Altar CA, Chen R, Nakamura T, Nakahashi T, Kambayashi J, Sun B, Tandon NN. Brain-derived neurotrophic factor is stored in human platelets and released by agonist stimulation. Thromb Haemost 2002; 87 (4): 728-734.

25. Lommatzsch M, Zingler D, Schuhbaeck K, Schloetcke K, Zingler C, Schuff-Werner P, Virchow JC. The impact of age, weight and gender on BDNF levels in human platelets and plasma. Neurobiol Aging 2005; 26 (1): 115-123.

26. Pan W, Banks WA, Fasold MB, Bluth J, Kastin AJ. Transport of brain-derived neurotrophic factor across the blood-brain barrier. Neuropharmacology 1998; 37 (12): 1553-1561.

27. Karege F, Perret G, Bondolfi G, Schwald M, Bertschy G, Aubry JM. Decreased serum brain-derived neurotrophic factor levels in major depressed patients. Psychiatry Res 2002; 109 (2): 143-148.

28. Shimizu E, Hashimoto K, Okamura N, Koike K, Komatsu N, Kumakiri C, Nakazato M, Watanabe H, Shinoda N, Okada S, Iyo M. Alterations of serum levels of brain-derived neurotrophic factor (BDNF) in depressed patients with or without antidepressants. Biol Psychiatry 2003; 54 (1): 70-75.

29. Gervasoni N, Aubry JM, Bondolfi G, Osiek C, Schwald M, Bertschy G, Karege F. Partial normalization of serum brain-derived neurotrophic factor in remitted patients after a major depressive episode. Neuropsychobiology 2005; 51 (4): 234-238.

30. Gonul AS, Akdeniz F, Taneli F, Donat O, Eker C, Vahip S. Effect of treatment on serum brain-derived neurotrophic factor levels in depressed patients. Eur Arch Psychiatry Clin Neurosci 2005; 255 (6): 381-386. 

31. Aydemir O, Deveci A, Taneli F. The effect of chronic anti-depressant treatment on serum brain-derived neurotrophic factor levels in depressed patients: a preliminary study. Prog Neuropsychopharmacol Biol Psychiatry 2005; 29 (2): 261-265.

32. Karege F, Bondolfi G, Gervasoni N, Schwald M, Aubry JM, Bertschy G. Low brain-derived neurotrophic factor (BDNF) levels in serum of depressed patients probably results from lowered platelet BDNF release unrelated to platelet reactivity. Biol Psychiatry 2005; 57 (9): 1068-1072.

33. American Psychiatric Association. Diagnostic and statistical manual of mental disorders. 4th ed. Washington DC, American Psychiatric Press, 1994.

34. Hamilton M. Development of a rating scale for primary depressive illness. Br J Soc Clin Psychol 1967; 6 (4): 278-296.

35. Hamilton M. A rating scale for depression. J Neurol Neurosurg Psychiatry 1960; 23: 56-62.

36. Yi JS, Bae SO, Ahn YM, Park DB, Noh KS, Shin HK, Woo HW, Lee HS, Han SI, Kim YS. Validity and reliability of the Korean Version of the Hamilton Depression Rating Scale (K-HDRS). J Korean Neuropsychiatr Assoc 2005; 44 (4): 456-465.

37. First M, Spitzer R, Gibbon M, Williams L. Structured Clinical Interview for DSM-IV Axis I Disorders (SCID). Washington DC, American Psychiatric Press, 1997.

38. Folstein M, Folstein S, McHugh P. "Mini-Mental State"; A practical method for grading the cognitive state of patients for the clinician. J Psychiatr Res 1975; 12 (3): 189-198.

39. Kang YW, Na DL, Hahn SH. A validity study on the Korean Mini-Mental State Examination (K-MMSE) in dementia patients. Korean J Neurology 1997; 15 (2): 300-308.

40. Beck AT, Ward CH, Mendelson M, Mock J, Erbaugh J. An inventory for measuring depression. Arch Gen Psychiatry 1961; 4: 561-571.

41. Han HM, Yeum TH, Shin YW, Kim KH, Yoon DJ, Jung GC. A Study on the standardization of Beck Depression Inventory-Centered on Normal Population (I). J Korean Neuropsychiatr Assoc 1986; 25: 487-502.

42. Delay J, Pichot P, Lelord G, Perse J. Validity of the Minnesota Multiphasic Personality Inventory. Ann Med Psychol (Paris) 1957; 115 (1): 90-107.

43. Kim YH, Kim JH, Kim JS, Noh MR, Yeum TH, Oh SW. Practice guideline for revised Minnesota Multiphasic Personality Inventory. Korea Guidance Ltd., 1989.

44. Radka SF, Holst PA, Fritsche M, Altar CA. Presence of brain-derived neurotrophic factor in brain and human and rat but not mouse serum detected by a sensitive and specific immunoassay. Brain Res 1996; 709 (1): 122-301.

45. Smith MA, Makino S, Kvetnansky R, Post RM. Effects of stress on neurotrophic factor expression in the rat brain. Ann N Y Acad Sci 1995; 771: 234-239.

46. Neeper SA, Gomez-Pinilla F, Choi J, Cotman CW. Physical activity increases mRNA for brain-derived neu-rotrophic factor and nerve growth factor in rat brain. Brain res 1996; 726 (1-2): 49-56.

47. Mattson MP, Duan W, Guo Z. Meal size and frequency affect neuronal plasticity and vulnerability to disease: cellular and molecular mechanisms. J Neurochem 2003; 84 (3): 417-431.