Skip to main content
Browse Subject Areas

Click through the PLOS taxonomy to find articles in your field.

For more information about PLOS Subject Areas, click here.

  • Loading metrics

Brain-Derived Neurotrophic Factor and Antidepressive Effect of Electroconvulsive Therapy: Systematic Review and Meta-Analyses of the Preclinical and Clinical Literature

  • M. Polyakova , (MP);

    Affiliations Max Planck Institute for Human Cognitive and Brain Sciences & Clinic for Cognitive Neurology, University Hospital, Leipzig, Germany, University Hospital Leipzig, Department of Psychiatry and Psychotherapy, Leipzig, Germany

  • M. L. Schroeter,

    Affiliation Max Planck Institute for Human Cognitive and Brain Sciences & Clinic for Cognitive Neurology, University Hospital, Leipzig, Germany

  • B. M. Elzinga,

    Affiliation Institute of Psychology, Leiden University and Leiden Institute for Brain and Cognition, Leiden University Medical Center, Leiden, The Netherlands

  • S. Holiga,

    Affiliation Max Planck Institute for Human Cognitive and Brain Sciences & Clinic for Cognitive Neurology, University Hospital, Leipzig, Germany

  • P. Schoenknecht,

    Affiliation University Hospital Leipzig, Department of Psychiatry and Psychotherapy, Leipzig, Germany

  • E. R. de Kloet,

    Affiliation Division of Medical Pharmacology, Division of Endocrinology, and Leiden Academic Center for Drug Research, Leiden University Medical Center, Leiden, The Netherlands

  • M. L. Molendijk (MP);

    Affiliation Institute of Psychology, Leiden University and Leiden Institute for Brain and Cognition, Leiden University Medical Center, Leiden, The Netherlands


Emerging data suggest that Electro-Convulsive Treatment (ECT) may reduce depressive symptoms by increasing the expression of Brain-Derived Neurotrophic Factor (BDNF). Yet, conflicting findings have been reported. For this reason we performed a systematic review and meta-analysis of the preclinical and clinical literature on the association between ECT treatment (ECS in animals) and changes in BDNF concentrations and their effect on behavior. In addition, regional brain expression of BDNF in mouse and human brains were compared using Allen Brain Atlas. ECS, over sham, increased BDNF mRNA and protein in animal brain (effect size [Hedge’s g]: 0.38―0.54; 258 effect-size estimates, N = 4,284) but not in serum (g = 0.06, 95% CI = -0.05―0.17). In humans, plasma but not serum BDNF increased following ECT (g = 0.72 vs. g = 0.14; 23 effect sizes, n = 281). The gradient of the BDNF increment in animal brains corresponded to the gradient of the BDNF gene expression according to the Allen brain atlas. Effect-size estimates were larger following more ECT sessions in animals (r = 0.37, P < .0001) and in humans (r = 0.55; P = 0.05). There were some indications that the increase in BDNF expression was associated with behavioral changes in rodents, but not in humans. We conclude that ECS in rodents and ECT in humans increase BDNF concentrations but this is not consistently associated with changes in behavior.


Electro Convulsive Treatment (ECT) has been used as a treatment for mood disorders for years. There is little doubt on the clinical efficacy of ECT [1, 2], yet, how it improves mood remains unclear [3, 4]. Emerging data have led to the idea that ECT may reduce depressive symptoms by increasing the expression of Brain-Derived Neurotrophic Factor (BDNF), a key regulator of neuronal functioning [5]. This idea rests on the neurotrophin hypothesis, which posits that depressive disorders are secondary to a stress-induced lowered expression of BDNF [6]. Complementary, it predicts that antidepressants are efficacious, because they increase BDNF expression and herewith boost neuronal plasticity [79].

Preclinical and clinical studies both have provided support for the neurotrophin hypothesis. Nibuya et al. [10], for instance, showed in rats that Electro-Convulsive Shocks (ECS, the equivalent of ECT in animals) increases the expression of hippocampal BDNF mRNA. This has been replicated and extended to other brain regions (e.g., the amygdala [11]) and was shown for BDNF protein levels [12]. Interestingly, and in line with the neurotrophin hypothesis, some studies show that the increase in BDNF following ECS is associated with a decrease in depression-like behaviors.

Measurements in brain tissue, as they are applied in preclinical studies, obviously cannot be pursued in humans. Clinical studies usually measure the change in peripheral (e.g., blood serum) BDNF protein concentrations over treatment with ECT. The validity of this approach is based on the observation that the brain is in part the source of BDNF in peripheral tissues [12, 13]. Clinical studies show that peripheral BDNF concentrations increase following treatment with ECT, as evidenced by a recent a meta-analysis (Hedge’s g = 0.38, 11 studies, 221 subjects) [14]. In contrast to some individual preclinical (e.g., Li et al. [15]) and clinical studies (e.g., Hu et al. [16]), this meta-analysis did not find evidence for the notion that changes in BDNF concentrations over treatment are related to the clinical efficacy of ECT. This omission may be due to a limited number of trials and patients and the use of group-level statistics [17]. An additional factor explaining the lack of association may be that serum and plasma BDNF measurement were merged in the analyses. Plasma levels are likely to reflect momentary BDNF protein expression, while serum levels reflect accumulated (over a period of about 10 days) BDNF [1820]. The combination of plasma and serum measurement in a single meta-analysis, as was done by Brunoni et al. [14], therefore may not be biologically plausible.

Notwithstanding some contradictory findings, the data above suggest a relation between ECT treatment and BDNF expression. The goal of this study, then, was to evaluate, through systematic review and meta-analyses, the preclinical (i.e., rodent) and clinical (i.e., human) literature on changes in BDNF concentrations and behavior over the course of ECS and ECT respectively. To fulfill this translational aim, we first will pool the preclinical literature on the relationships between ECS, BDNF and depression-like behavior. Next, we will aggregate effect-sizes of ECT treatment on BDNF concentrations and clinical improvement as they are reported in the human literature. This will be done partially using meta-analysis on individual data because this better suits the questions at hand given a limited number of trials and patients that are available [17].

Materials and Methods

We adhered to the guidelines that are recommended by the preferred reporting items for systematic reviews and meta-analyses statement [21].

Search Strategy

We searched PUBMED, Embase, and PsychInfo through December 1st 2014 to identify eligible studies on changes in peripheral and central BDNF concentrations as a function of treatment with ECT. The following keywords were used: ‘electroconvulsive’ or ‘ECT’ or ‘ECS’ in combination with ‘BDNF’ or ‘brain derived neurotrophic factor’. This was supplemented by backward searches in which the references to the seminal papers of interest were screened for preclinical and clinical studies and by examining the reference sections of the retrieved papers. The literature search, decisions on inclusion, data extraction, and quality control were performed independently by two of the authors (MP and MM).

Inclusion and Exclusion Criteria

We included peer-reviewed preclinical and clinical studies that reported data on BDNF concentrations as a function of ECS/ECT (i.e., ECS/ECT versus sham and pre versus post treatment). Inclusion was independent of ECS/ECT characteristics (e.g., number of sessions) and methodological characteristics of the study (e.g., tissue in which BDNF was sampled). For the clinical studies diagnosis of major- or bipolar depression had to be based on international classifications.

Non-empirical studies such as reviews were excluded according to review protocol, as were case studies, studies that were not peer reviewed, and studies that were not written in Dutch, English, French, German or Spanish. Where study samples overlapped we excluded the study that reported on the fewest number of subjects.

Data Extraction

From each paper we extracted, as primary outcomes, mean BDNF concentrations (and Standard Deviation [SD]) in treatment conditions versus sham and/or before and after ECS/ECT or indices on this change (e.g., the standardized mean difference). We also extracted data on mean age, gender distribution, specifics of the ECS/ECT treatment, and the method that was applied to quantify the amount of BDNF (e.g., RT-PCR).

From the preclinical studies we further extracted data on the strain of animal that was used, the weight and age of the animals, the brain-region in which BDNF was measured, and the amount of time between ECS treatment and decapitation. Data on behavioral changes due to ECS were extracted where provided.

From the clinical studies we in addition extracted data on depression severity pre- and post ECT, whether participants exhibited a clinical response to ECT, the antidepressant that were used, and the amount of time between the last ECT session and blood draw for BDNF determination. In order to perform subgroup comparisons according to treatment response we contacted the authors of the clinical studies and asked them to provide anonymised Individual Patients Data (IPD) [17]. In those cases where non-significant results were reported (e.g., P>.05) and authors did not reply to our request for exact outcome data, we set the association at P = .50, indicating no association.

We assessed the methodological quality of the preclinical and clinical studies using the ARRIVE guidelines[22] and the Newcastle-Ottawa Scale (NOS) [23] respectively. In addition we used the risk of bias assessment tool for the longitudinal studies [24]. We refer to the Supplement for detailed information on quality assessment (S1 Text, S1 Table, S2 Table, S3 Table).

Statistical Analysis

Analyses were performed using Comprehensive Meta-Analyses 2.0 [25] and SPSS version 21.0 [26].Random effects models (i.e., models that include sampling- and study level error) were applied to calculate pooled effect-sizes on changes in central and peripheral BDNF concentrations as a function of ECS/ECT. As effect-size measure we chose to use Hedges’ g, a standardized metric that corrects for bias related to small sample sizes [27]. All outcomes were weighted using inverse variance methods [25]. Statistical significance was assessed using a Z-statistic at a Confidence Interval (CI) of 95%. The amount of between-study heterogeneity in outcomes was quantified using the I2 measure [28] and assessed for statistical significance using the Q-statistic[25].

The stability of our results was evaluated through meta-analyses that were run in specific subgroups: (I) by brain region in which BDNF was assessed (clustered as follows: Dentate Gyrus [DG], hippocampus not DG, cortex, other brain regions, and in serum [S4 Table], (II) single versus multiple ECT sessions, and (III) the type of BDNF that was measured (i.e., BDNF mRNA versus protein and BDNF in serum versus in plasma). The possible moderating effects of between-study differences on outcomes were evaluated by calculating correlation coefficients between the values for the moderator and the outcome of the studies.

For the analyses on preclinical data, the animal strain that was used, duration of treatment, the amount of time between the last ECT session and decapitation for BDNF measurements, and the quality score were considered as potential moderators. For clinical data analysis, obtained IPD were combined with the aggregated data using a two-step approach. In a first step summary statistics were calculated for each subgroup from single studies. In the second step summary statistics from the IPD were combined in meta-analysis as described above. Treatment response was considered as reduction of depression severity scores by ≥50%. Duration of treatment and the quality score were considered as potential moderators of the effect-sizes retrieved from clinical studies.

Visual inspection of funnel plots and the Egger test were used to assess publication bias [29]. In case of publication bias we used trim-and-fill procedures to estimate effect-sizes after bias has been taken into account [30].


Preclinical Studies

Our search generated 97 papers of which 23 [1011,15,3351] fulfilled the inclusion criteria (see Fig 1 for a flow-chart). From these we could extract 280 effect-size estimates (k) on a total of 4,670 animals (mean n = 17 per effect-size, range 8–30) on changes in BDNF concentrations in animals that were subjected to ECS as compared to sham treatment or, in one case, to baseline.[31] Mean number of ECS sessions was 5 (range: 1–14). Mean time that passed between last ECS session and decapitation was 40 hours (range: 1–504 hours). We refer to Table 1 for the included studies and general information on them. S5 Table and S6 Table provide additional information on the animals that were used and the methods that were applied.

Fig 1. Prisma flow diagram of the search strategy and its results.

Table 1. Basic information on the preclinical studies included in the meta-analysis.

Meta-analysis over preclinical findings

ECS was associated with increased BDNF concentrations in comparison to sham treatment (g = 0.40, 95% CI = 0.35―0.44, P < .0001; 280 effect-sizes, N = 4,284). Meta-analyses by specific brain region showed a larger effect-size (P < .05) when BDNF was assessed in the DG (g = 0.54) as compared to assessments in the hippocampus and the cortex (g: 0.38―0.41 respectively). Yet, effect-sizes were significant regardless in which brain area BDNF was sampled (see Table 2). Interestingly, the observed gradient of ECS induced increases in BDNF protein corresponds to the gradient of BDNF gene expression across the whole brain in mice and humans as assessed in the genome wide atlas of the Allen Institute for Brain Sciences (Seattle, WA, USA, see [52]. Results of this analysis are illustrated in Fig 2 with highest gene expression in DG, followed by hippocampus and other brain regions.

Fig 2. Gene expression of BDNF across the whole brain as assessed in Allen Brain Atlas (Seattle, WA, USA).

In mice, gene expression was the highest in the DG/hippocampus as investigated by In Situ Hybridization (ISH), where warm colors indicate high expression. Note that contrast and brightness were enhanced in original images to increase visibility of the effects here. Black bars correspond to 2 mm. Gene expression in humans is shown as individually normalized gene expression (Z-scores normalized to whole human brain expression). The heat map shows scores across the whole human brain and for each of the six subjects contained in the database beside each other, where red indicates high and green indicates low expression. Bars represent mean normalized gene expression and standard deviation across one female and five male subjects included. Search conducted on 17th October 2014 for human and on 9th October 2014 for mice data. *** P < 0.001; ** P < 0.01; * P < 0.05, two-tailed Student’s t-test against 0. For details on the methods we refer to Mueller et al.[53]

Table 2. Pooled effect-size estimates, heterogeneity and publication bias for the animal studies by sub-group meta-analyses indicated in the row.

Evidence for increases in serum BDNF concentrations (i.e., in blood serum) following ECS was not found in the preclinical data (g = 0.06, 95% CI = -0.05―0.17). In fact, the pooled effect-size on serum measurement was smaller as compared to the ones calculated on central BDNF (P-values all < .001). Studies that subjected animals to multiple ECS’s yielded larger effect-size as compared to studies that applied single ECS (P < .0001). Yet, also a single ECS session was associated with an increase in BDNF (see Table 2). The pooled effect size that was derived from studies that measured BDNF mRNA was larger as the one from studies that measured BDNF protein (P < .0001) although the latter also was statistically significant (see Table 2).

Between-study heterogeneity in outcomes was identified (I2 = 51%, Q = 572.13, P < .00001). The number of ECS sessions that was applied and the time that passed between the last ECS session and measurement appeared to be sources of the observed heterogeneity. A larger number of treatment sessions, in general was associated with larger effect size estimates (r = 0.36, R2 = 0.13, P < .0001) and a longer gap in time between the last ECS session and decapitation with smaller effect-size-estimates (r = -0.27, R2 = 0.07, P < .0001). The correlation between the number of ECS and ECS induced increase in BDNF was also present within the multiple treated animals (r = 0.35, R2 = 0.13, [202 data points], P < .0001). The methodological quality of a study was unrelated to outcome.

The funnel plot and the Egger’s test suggested evidence for publication bias in the overall analysis (t[278] = 10.41, P < .0001). Imputation of 15, presumed missing, effect-size estimates resulted in a symmetric funnel-plot. The pooled effect-size estimate that was recalculated after imputation was only slightly smaller as compared to the one derived in the original analysis (g = 0.38, 95% CI = 0.35―0.41). Between-study heterogeneity, correlations between outcomes and moderators, and publication bias in outcomes were rather similar in analyses that were run separately in the subgroups (see Table 3).

Table 3. Pearson’s correlation coefficients on the relation between study characteristics and study effect size (by meta-analysis indicated in the columns).

ECS, BDNF and changes in behaviour.

There was too little comparable data on behavioral tests (e.g., the open-field test) to perform meta-analysis on. In case similar behavioral paradigms were applied, often the outcome measures over studies were different. This was for instance so for swimming time in the FST for which we could extract 48 effect-size estimates (910 animals, range: 11―30 per effect-size) on total swimming time in the FST. Together, these showed that ECS, over sham, increased swimming time (g = 0.26, 95% CI = 0.19―0.32, P < .0001). The increase in swimming time correlated positively with the increase in BDNF (r = 0.37, R2 = 0.14, P < .001). Note though that these effect sizes came from only four studies that widely differed in for instance in time of sacrifice after ECS and other variables that potentially can confound the observed relation. The association, thus, should be interpreted with caution.

Sensitivity analyses.

None of the study findings was unduly driven by the effect of a particular study (data not shown). Furthermore, effect-size estimates were not related (P = .49) to whether or a particular study used a stress paradigm (e.g., chronic unpredictable mild stress). Method of BDNF measurement was not associated with the amount of change in detectable BDNF (P = .17; see S6 Table for the methods of measurement by study). Animal strain was tested as a potential effect modifier (see S6 Table for the animal strain that was used in each individual study). Analyses showed that there were no differences in ECS induced increases in BDNF as a function of strain of animal that was used in the experiment (P = .18).

Clinical Studies

Our search for clinical studies generated 111 publications of which 14 fulfilled the inclusion criteria (see Fig 1 for a flow-chart). From these papers we obtained 23 effect-size estimates on changes in BDNF concentrations over the course of ECT (N = 250 subjects a [mean n = 13 per effect-size, range 3–48]). Ten studies [16, 5462] reported on serum BDNF alternations (16 effect sizes) and 4 studies [6366] (7 effect sizes) on plasma BDNF alterations. Table 4 and S7 Table provide details of the included studies.

Table 4. Basic information on the clinical studies included in the meta-analysis.

Meta-analysis over clinical findings.

Overall peripheral BDNF was significantly increased after ECT as compared to baseline (g = 0.35, 95% CI = 0.034–0.67, P = 0.03; 23 effect sizes, n = 281). BDNF levels increased in plasma (g = 0.72, 95% CI = 0.22–1.23, P = 0.004; 7 effect sizes, n = 108) but not in serum (g = 0.14, 95% CI = -0.29–0.56, P = 0.67; 16 effect sizes, n = 173). However, the difference between serum and plasma subgroups did not rich the significance threshold (P = 0.10; Table 5). When subdivided into responders and non-responders subgroups, BDNF increased non-significantly in both the responders- (g = 0.40 95% CI = 0.02–0.82, P = 0.06; 13 effect sizes, n = 214) and non-responders subgroups (g = 0.22 95% CI = -0.38–0.82, P = 0.48; 9 effect sizes, n = 47). There was no different pattern of results when comparing the pooled effect sizes from studies that measured BDNF in serum versus plasma (Table 5). However, significant differences could be observed between plasma and serum BDNF in the non-responders subgroups (P = 0.05).

Table 5. Pooled effect-size estimates, heterogeneity and publication bias for the clinical studies by sub-group meta-analyses indicated in the row.

Sensitivity analysis showed that the results were not substantially affected by a single study. We observed overall high heterogeneity in outcomes between the studies (Q = 63.11[22] P<0.001, I2 = 65.14%). This appeared to be driven by the responder subgroups in both serum and plasma (see Table 5). In line with the rodent findings, the number of ECT treatments was positively correlated with the effect sizes in combined serum and plasma subgroup (r = 0.55; P = 0.05). The number of subjects and methodological quality of the study was not associated with outcomes (data not shown).

Publication bias was detected in the serum subgroup and induced by the studies by Stelzhammer et al. [59] and Kleimann et al. [62], two negative studies with particularly low power. Correction for publication bias by Trim-and-Fill procedure led to an increased effect size (g = 0.57, 95% CI = 0.11―1.04; 11 effect sizes, n = 101). Overall, and in the plasma subgroup, no publication bias was detected.


Our systematic review and meta-analyses investigated changes in BDNF concentrations as a function of ECS and ECT. Our main findings are: (A) in rodents, ECS increases BDNF mRNA and protein concentration (or synthesis/release) in the brain, with largest effect sizes measured in the DG, (B) the increase in BDNF is positively correlated with number of treatments and negatively with the time between the last ECT and BDNF measurement, (C) BDNF concentrations do not increase in the course of treatment in rodent and human serum, yet they increased in human plasma, and (D) the increase in BDNF following ECT is also related to the number of treatment sessions but not to clinical outcome in human studies.

In preclinical studies ECS increased BDNF secretion throughout the brain. Activation of distinct promoters of the BDNF gene is responsible for a differently regulated BDNF expression over brain regions [67, 68]. Four out of nine possible BDNF transcripts are expressed in the rat brain [69]. While in most brain regions one or two transcripts are produced, all four are activated in the DG following ECT [36]. Interestingly, BDNF expression as elicited by ECS appeared to be highest in the DG. This relates well to what the Allen brain atlas shows: BDNF expression in the DG of mice and human brains is highest in this region. A constant supply of BDNF here is not restricted to effects of ECS. This may serve neurogenesis, as the DG is one of the main sources of progenitor cells [70, 71].

The effect of single ECS on BDNF concentrations seems to be short-lived (6–8 hours [10, 49]) and does not involve a hard reset after which BDNF expression remains at a constant higher level. The effect of multiple ECS lasted longer as compared to single treatment: up to 14 days post-ECS [15, 31, 39, 45]. On the meta-analytical level this was reflected by a positive correlation between number of treatments and BDNF levels in rodents, and a trend towards such an association (P = .06) in humans.

Interestingly, effect sizes were larger for BDNF mRNA as compared to protein concentrations. Several posttranscriptional mechanisms can be responsible for this. First, protein synthesis may be inhibited by a specific class of microRNA molecules, that bind target mRNA and induce its degradation. Several microRNAs are associated with BDNF depletion [72, 73], one of them, microRNA-212, was increased after ECS in rat’s DG [46]. Second, there is evidence of activity-dependent mRNA trafficking of BDNF to dendrites, where it can be stored and translated on demand [74]. Third, an increased BDNF turnover after ECS was proposed[38] and makes sense in light of findings of neurogenesis after ECT [71, 75].

Once BDNF is synthesized it can act locally, be transferred to neighboring brain areas through axonal anterograde transport or secreted to the blood stream. The later property allowed scientists to make inferences about central BDNF secretion from peripheral measurements. However, initial findings of a high positive correlation between central and serum BDNF [76] were not confirmed [77, 78] or at least depended on animal strain and brain region [31]. Neither a correlation between CSF and serum BDNF in humans was demonstrated [79]. In rodents we demonstrated increments in brain but not in the serum BDNF levels.

In clinical studies, ECT increased peripheral BDNF levels with a small to moderate effect size (g = 0.35). Compared to a previous meta-analysis on this topic [14], we included newly published studies, obtained individual patient data and took the source of BDNF (i.e., plasma versus serum) into account. This approach revealed significantly enhanced BDNF after ECT in plasma and not in serum.

Although both plasma and serum BDNF levels are decreased in acute major and bipolar depression [24, 80], they seem to restore differently following antidepressant treatment [24]. The difference between responders and non-responders that we observed in serum BDNF after pharmacological antidepressant treatment was not demonstrated after ECT. Neither did we observe an increase in serum BDNF after ECT. This differs for plasma measurements, where ECT seems to lead to an increase of BDNF but antidepressant treatment did not [24]. Such difference may point to different mechanisms of action of ECT and antidepressants on BDNF synthesis and release.


Our study has a number of limitations. First, obviously we could not match the preclinical and clinical studies according to depressive state, only (roughly) according to the treatment applied. Most of the animal studies used healthy male animals and did not account for the effects of sex and disease on BDNF. The clinical studies, in turn, included both sexes and were based on treatment-resistant depression cases. Furthermore, none of these studies controlled for relevant confounders in longitudinal studies assessing BDNF, such as seasonality [81]. Plasma BDNF studies could be further confounded by measurement errors [24]. Secondly, due to limited data we had to combine treatment effects on major- and bipolar depression even though imaging studies show differential response to ECT for these two groups [82]. Thirdly, most of the clinical studies included antidepressants and ECT premedication which may have affected BDNF concentrations. Fourthly, while meta-analysis of preclinical data had enough power and showed small to medium heterogeneity, the meta-analysis of clinical data was underpowered and showed signs of publication bias. Due to the limited power we could not control for the impact of ELISA kit manufacturer on effect sizes. Finally, effect-size estimates for the preclinical data may have been suboptimal in terms of precision because they were often estimated based on P-value and N.


Despite the limitations, animal and human studies seem to complement each other with regard to effects of ECT on BDNF: ECT increases brain BDNF in animals and plasma BDNF in humans. In animals regional BDNF increments after ECT (i.e., the DG) corresponded to areas with distinct expression shown in the Allen brain atlas. Besides, multiple treatments as compared to single ECT were associated with a larger increase in BDNF in both animals and humans, which is suggestive for a dose-response effect of ECS on BDNF.

Future Directions

The questions that remain unsolved are: (1) why plasma but not serum BDNF increased in human studies, (2) what is the relationship between BDNF and behavior, and (3) are increments in BDNF after ECT/ECS related to neurogenesis?

The potential differences between serum and plasma may arise from several aspects. Firstly, plasma BDNF levels reflect momentary BDNF content whereas serum levels reflect BDNF that has been accumulated over several days or even weeks [1820]. Secondly, plasma measurements are very sensitive to the laboratory conditions and, thus, error prone [24]. Future studies (following strict methodological recommendations) should clarify whether plasma increment is not an artifact and further investigate the nature of plasma and serum BDNF.

A larger number of studies is needed to understand the relation of behavioral outcomes to BDNF levels. For clinical studies such outcome measurement is well established: response to treatment or clinical remission. Preclinical studies, however, reported rather different, in terms of timeframe and behavioral assessment, data. Therefore, for the later at least partial overlap in outcome variables with previous studies is needed.

Though the behavioral data is still mixed, neurogenesis is required to achieve antidepressive effect of ECS [83]. Survival of newborn neurons is supported by BDNF [84]. The causality and the dose-response relationships between ECS, BDNF, neurogenesis and behavior are the next questions to adress. Moderators of BDNF functioning, most notably the common genetic variant val66met that has been associated with activity dependent BDNF expression [85], might be considered. Relating variation at this locus to hippocampal morphology [86] and functioning [87] however has thus far led to mixed results.

Supporting Information

S1 Table. Quality of the included preclinical studies.


S2 Table. Quality of the included clinical studies as measured with the Newcastle-Ottawa Scale.


S3 Table. Quality of the included clinical studies as measured with the RBLS.


S4 Table. Brain regions in which BDNF was sampled in the preclinical studies that we included.


S5 Table. Basic information on the animals that were used in the preclinical studies that were included in our meta-analysis.


S6 Table. Basic methodological information on the preclinical studies that were included in our meta-analysis.


S7 Table. Basic information on the patients that were included in the clinical studies that were included in our meta-analysis.


S1 Text. Quality assessment of the included studies.


Author Contributions

Conceived and designed the experiments: MP MLS BME SH PS ERK MM. Performed the experiments: MM MP SH MLS. Analyzed the data: MM MP SH MLS. Wrote the paper: MP MLS BME SH PS ERK MM.


  1. 1. Kho KH, van Vreeswijk MF, Simpson S, Zwinderman AH. A meta-analysis of electroconvulsive therapy efficacy in depression. The journal of ECT. 2003;19(3):139–47. Epub 2003/09/16. pmid:12972983.
  2. 2. Dierckx B, Heijnen WT, van den Broek WW, Birkenhager TK. Efficacy of electroconvulsive therapy in bipolar versus unipolar major depression: a meta-analysis. Bipolar Disord. 2012;14(2):146–50. Epub 2012/03/17. pmid:22420590.
  3. 3. Weiner RD, Falcone G. Electroconvulsive therapy: how effective is it? Journal of the American Psychiatric Nurses Association. 2011;17(3):217–8. Epub 2011/06/10. pmid:21653492.
  4. 4. Guloksuz S, Rutten BP, Arts B, van Os J, Kenis G. The immune system and electroconvulsive therapy for depression. The journal of ECT. 2014;30(2):132–7. Epub 2014/04/24. pmid:24755720.
  5. 5. Bouckaert F, Sienaert P, Obbels J, Dols A, Vandenbulcke M, Stek M, et al. ECT: its brain enabling effects: a review of electroconvulsive therapy-induced structural brain plasticity. The journal of ECT. 2014;30(2):143–51. Epub 2014/05/09. pmid:24810772.
  6. 6. Li N, Liu GT. The novel squamosamide derivative FLZ enhances BDNF/TrkB/CREB signaling and inhibits neuronal apoptosis in APP/PS1 mice. Acta Pharmacol Sin. 2010;31(3):265–72. Epub 2010/02/16. pmid:20154710; PubMed Central PMCID: PMCPmc4002416.
  7. 7. Duman RS, Heninger GR, Nestler EJ. A molecular and cellular theory of depression. Archives of General Psychiatry. 1997;54(7):597–606. pmid:WOS:A1997XK68700001.
  8. 8. Duman RS, Monteggia LM. A neurotrophic model for stress-related mood disorders. Biological Psychiatry. 2006;59(12):1116–27. pmid:WOS:000238692300003.
  9. 9. Park H, Poo MM. Neurotrophin regulation of neural circuit development and function. Nature reviews Neuroscience. 2013;14(1):7–23. Epub 2012/12/21. pmid:23254191.
  10. 10. Nibuya M, Morinobu S, Duman RS. Regulation of BDNF and trkB mRNA in rat brain by chronic electroconvulsive seizure and antidepressant drug treatments. J Neurosci. 1995;15(11):7539–47. Epub 1995/11/01. pmid:7472505.
  11. 11. Altar CA, Whitehead RE, Chen RY, Wortwein G, Madsen TM. Effects of electroconvulsive seizures and antidepressant drugs on brain-derived neurotrophic factor protein in rat brain. Biological Psychiatry. 2003;54(7):703–9. pmid:WOS:000185500500006.
  12. 12. 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):1553–61. Epub 1999/01/14. pmid:9886678.
  13. 13. Klein AB, Trajkovska V, Erritzoe D, Haugbol S, Madsen J, Baare W, et al. Cerebral 5-HT2A receptor and serotonin transporter binding in humans are not affected by the val66met BDNF polymorphism status or blood BDNF levels. Journal of Cerebral Blood Flow and Metabolism. 2010;30(11):E1–E7. pmid:WOS:000283736300001.
  14. 14. Brunoni AR, Baeken C, Machado-Vieira R, Gattaz WF, Vanderhasselt M- A. BDNF blood levels after electroconvulsive therapy in patients with mood disorders: A systematic review and meta-analysis. The world journal of biological psychiatry: the official journal of the World Federation of Societies of Biological Psychiatry. 2014;15(5):411–8. pmid:MEDLINE:24628093.
  15. 15. Li B, Suemaru K, Cui R, Araki H. Repeated electroconvulsive stimuli have long-lasting effects on hippocampal BDNF and decrease immobility time in the rat forced swim test. Life Sci. 2007;80(16):1539–43. Epub 2007/02/20. pmid:17306836.
  16. 16. Hu Y, Yu X, Yang F, Si T, Wang W, Tan Y, et al. The level of serum brain-derived neurotrophic factor is associated with the therapeutic efficacy of modified electroconvulsive therapy in Chinese patients with depression. The journal of ECT. 2010;26(2):121–5. Epub 2009/11/26. pmid:19935088.
  17. 17. Cooper H, Patall EA. The relative benefits of meta-analysis conducted with individual participant data versus aggregated data. Psychological methods. 2009;14(2):165–76. Epub 2009/06/03. pmid:19485627.
  18. 18. Fujimura H, Altar CA, Chen R, Nakamura T, Nakahashi T, Kambayashi J, et al. Brain-derived neurotrophic factor is stored in human platelets and released by agonist stimulation. Thrombosis and haemostasis. 2002;87(4):728–34. Epub 2002/05/15. pmid:12008958.
  19. 19. Harker LA, Roskos LK, Marzec UM, Carter RA, Cherry JK, Sundell B, et al. Effects of megakaryocyte growth and development factor on platelet production, platelet life span, and platelet function in healthy human volunteers. Blood. 2000;95(8):2514–22. Epub 2001/02/07. pmid:10753829.
  20. 20. 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):122–30. pmid:WOS:A1996TX90900015.
  21. 21. Moher D, Liberati A, Tetzlaff J, Altman DG. Preferred reporting items for systematic reviews and meta-analyses: the PRISMA statement. Journal of clinical epidemiology. 2009;62(10):1006–12. Epub 2009/07/28. pmid:19631508.
  22. 22. Kilkenny C, Browne WJ, Cuthill IC, Emerson M, Altman DG. Improving bioscience research reporting: The ARRIVE guidelines for reporting animal research. Journal of pharmacology & pharmacotherapeutics. 2010;1(2):94–9. Epub 2011/02/26. pmid:21350617; PubMed Central PMCID: PMCPmc3043335.
  23. 23. Wells GA SB, O’Connel D. The Newcastle-Ottawa Scale (NOS) for assessing the quality non-randomised studies in meta-analyses.
  24. 24. Polyakova M, Stuke K, Schuemberg K, Mueller K, Schoenknecht P, Schroeter ML. BDNF as a biomarker for successful treatment of mood disorders: a systematic & quantitative meta-analysis. Journal of affective disorders. 2014;174c:432–40. Epub 2015/01/02. pmid:25553404.
  25. 25. Borenstein M HL, Higgins JPT. Comprehensive Meta Analysis. 2.2.064 ed. Englewood, NJ, USA: Wiley; 2009.
  26. 26. IBM SPSS Statistics for Windows, Version 21.0. Armonk, NY: IBM Corp: IBM Corp.; Released 2012.
  27. 27. Hedges LV, Olkin I. Statistical Methods for Meta-analysis: Academic Press; 1985.
  28. 28. Higgins JP, Thompson SG. Quantifying heterogeneity in a meta-analysis. Statistics in medicine. 2002;21(11):1539–58. Epub 2002/07/12. pmid:12111919.
  29. 29. Egger M, Davey Smith G, Schneider M, Minder C. Bias in meta-analysis detected by a simple, graphical test. BMJ (Clinical research ed). 1997;315(7109):629–34. Epub 1997/10/06. pmid:9310563; PubMed Central PMCID: PMCPmc2127453.
  30. 30. Peters JL, Sutton AJ, Jones DR, Abrams KR, Rushton L. Performance of the trim and fill method in the presence of publication bias and between-study heterogeneity. Statistics in medicine. 2007;26(25):4544–62. Epub 2007/05/04. pmid:17476644.
  31. 31. Sartorius A, Hellweg R, Litzke J, Vogt M, Dormann C, Vollmayr B, et al. Correlations and discrepancies between serum and brain tissue levels of neurotrophins after electroconvulsive treatment in rats. Pharmacopsychiatry. 2009;42(6):270–6. Epub 2009/11/20. pmid:19924587.
  32. 32. Lindefors N, Brodin E, Metsis M. Spatiotemporal selective effects on brain-derived neurotrophic factor and trkB messenger-rna in rat hippocampus by electroconvulsive shock. Neuroscience. 1995;65(3):661–70. pmid:WOS:A1995QN86100003.
  33. 33. Zetterstrom TS, Pei Q, Grahame-Smith DG. Repeated electroconvulsive shock extends the duration of enhanced gene expression for BDNF in rat brain compared with a single administration. Brain research Molecular brain research. 1998;57(1):106–10. pmid:MEDLINE:9630550.
  34. 34. Chen ACH, Shin KH, Duman RS, Sanacora G. ECS-induced mossy fiber sprouting and BDNF expression are attenuated by ketamine pretreatment. Journal of Ect. 2001;17(1):27–32. pmid:WOS:000167512900006.
  35. 35. Angelucci F, Aloe L, Jimenez-Vasquez P, Mathe AA. Electroconvulsive stimuli alter nerve growth factor but not brain-derived neurotrophic factor concentrations in brains of a rat model of depression. Neuropeptides. 2003;37(1):51–6. pmid:WOS:000182057800007.
  36. 36. Dias BG, Banerjee SB, Duman RS, Vaidya VA. Differential regulation of brain derived neurotrophic factor transcripts by antidepressant treatments in the adult rat brain. Neuropharmacology. 2003;45(4):553–63. pmid:WOS:000185048900012.
  37. 37. Newton SS, Collier EF, Hunsberger J, Adams D, Terwilliger R, Selvanayagam E, et al. Gene profile of electroconvulsive seizures: Induction of neurotrophic and angiogenic factors. Journal of Neuroscience. 2003;23(34):10841–51. pmid:WOS:000186902500012.
  38. 38. Jacobsen JPR, Mork A. The effect of escitalopram, desipramine, electroconvulsive seizures and lithium on brain-derived neurotrophic factor mRNA and protein expression in the rat brain and the correlation to 5-HT and 5-HIAA levels. Brain Res. 2004;1024(1–2):183–92. pmid:WOS:000224664800020.
  39. 39. Li B, Suemaru K, Cui R, Kitamura Y, Gomita Y, Araki H. Repeated electroconvulsive stimuli increase brain-derived neurotrophic factor in ACTH-treated rats. European Journal of Pharmacology. 2006;529(1–3):114–21. pmid:WOS:000234636400015.
  40. 40. Ploski JE, Newton SS, Duman RS. Electroconvulsive seizure-induced gene expression profile of the hippocampus dentate gyrus granule cell layer. Journal of Neurochemistry. 2006;99(4):1122–32. pmid:WOS:000241562900008.
  41. 41. Conti B, Maier R, Barr AM, Morale MC, Lu X, Sanna PP, et al. Region-specific transcriptional changes following the three antidepressant treatments electro convulsive therapy, sleep deprivation and fluoxetine. Molecular psychiatry. 2007;12(2):167–89. pmid:WOS:000243783700006.
  42. 42. Gersner R, Toth E, Isserles M, Zangen A. Site-specific antidepressant effects of repeated subconvulsive electrical stimulation: potential role of brain-derived neurotrophic factor. Biological Psychiatry. 2010;67(2):125–32. pmid:WOS:000273201800006.
  43. 43. Kyeremanteng C, James J, Mackay J, Merali Z. A study of brain and serum brain-derived neurotrophic factor protein in Wistar and Wistar-Kyoto rat strains after electroconvulsive stimulus. Pharmacopsychiatry. 2012;45(6):244–9. Epub 2012/03/29. pmid:22454252.
  44. 44. Luo J, Min S, Wei K, Zhang J, Liu Y. Propofol interacts with stimulus intensities of electroconvulsive shock to regulate behavior and hippocampal BDNF in a rat model of depression. Psychiatry research. 2012;198(2):300–6. pmid:WOS:000311987000022.
  45. 45. O'Donovan S, Kennedy M, Guinan B, O'Mara S, McLoughlin DM. A comparison of brief pulse and ultrabrief pulse electroconvulsive stimulation on rodent brain and behaviour. Prog Neuro-Psychopharmacol Biol Psychiatry. 2012;37(1):147–52. pmid:WOS:000302672300022.
  46. 46. Ryan KM, O'Donovan SM, McLoughlin DM. Electroconvulsive stimulation alters levels of BDNF-associated microRNAs. Neuroscience letters. 2013;549:125–9. pmid:WOS:000323140000025.
  47. 47. Segawa M, Morinobu S, Matsumoto T, Fuchikami M, Yamawaki S. Electroconvulsive seizure, but not imipramine, rapidly up-regulates pro-BDNF and t-PA, leading to mature BDNF production, in the rat hippocampus. International Journal of Neuropsychopharmacology. 2013;16(2):339–50. pmid:WOS:000314996100010.
  48. 48. Segi-Nishida E, Sukeno M, Imoto Y, Kira T, Sakaida M, Tsuchiya S, et al. Electroconvulsive seizures activate anorexigenic signals in the ventromedial nuclei of the hypothalamus. Neuropharmacology. 2013;71:164–73. pmid:WOS:000320424200017.
  49. 49. Dyrvig M, Christiansen SH, Woldbye DPD, Lichota J. Temporal gene expression profile after acute electroconvulsive stimulation in the rat. Gene. 2014;539(1):8–14. pmid:WOS:000333496500002.
  50. 50. Kyeremanteng C, MacKay JC, James JS, Kent P, Cayer C, Anisman H, et al. Effects of electroconvulsive seizures on depression-related behavior, memory and neurochemical changes in Wistar and Wistar-Kyoto rats. Prog Neuropsychopharmacol Biol Psychiatry. 2014;54:170–8. Epub 2014/05/30. pmid:24871798.
  51. 51. Gersner R, Toth E, Isserles M, Zangen A. Site-Specific Antidepressant Effects of Repeated Subconvulsive Electrical Stimulation: Potential Role of Brain-Derived Neurotrophic Factor. Biological Psychiatry. 2010;67(2):125–32. pmid:WOS:000273201800006.
  52. 52. Jones AR, Overly CC, Sunkin SM. The Allen Brain Atlas: 5 years and beyond. Nature reviews Neuroscience. 2009;10(11):821–8. Epub 2009/10/15. pmid:19826436.
  53. 53. Mueller K, Sacher J, Arelin K, Holiga S, Kratzsch J, Villringer A, et al. Overweight and obesity are associated with neuronal injury in the human cerebellum and hippocampus in young adults: a combined MRI, serum marker and gene expression study. Transl Psychiatry. 2012;2:e200. Epub 2012/12/06. pmid:23212584; PubMed Central PMCID: PMCPmc3565188.
  54. 54. Bocchio-Chiavetto L, Zanardini R, Bortolomasi M, Abate M, Segala M, Giacopuzzi M, et al. Electroconvulsive therapy (ECT) increases serum brain derived neurotrophic factor (BDNF) in drug resistant depressed patients. European Neuropsychopharmacology. 2006;16(8):620–4. pmid:WOS:000242695600010.
  55. 55. Okamoto T, Yoshimura R, Ikenouchi-Sugita A, Hori H, Umene-Nakano W, Inoue Y, et al. Efficacy of electroconvulsive therapy is associated with changing blood levels of homovanillic acid and brain-derived neurotrophic factor (BDNF) in refractory depressed patients: A pilot study. Prog Neuro-Psychopharmacol Biol Psychiatry. 2008;32(5):1185–90. pmid:WOS:000258213200010.
  56. 56. Fernandes B, Gama CS, Massuda R, Torres M, Camargo D, Kunz M, et al. Serum brain-derived neurotrophic factor (BDNF) is not associated with response to electroconvulsive therapy (ECT): a pilot study in drug resistant depressed patients. Neuroscience letters. 2009;453(3):195–8. Epub 2009/05/12. pmid:19429034.
  57. 57. Gronli O, Stensland GO, Wynn R, Olstad R. Neurotrophic factors in serum following ECT: a pilot study. The world journal of biological psychiatry: the official journal of the World Federation of Societies of Biological Psychiatry. 2009;10(4):295–301. Epub 2009/11/20. pmid:19921971.
  58. 58. Gedge L, Beaudoin A, Lazowski L, du Toit R, Jokic R, Milev R. Effects of electroconvulsive therapy and repetitive transcranial magnetic stimulation on serum brain-derived neurotrophic factor levels in patients with depression. Front Psychiatry. 2012;3:12. Epub 2012/03/01. pmid:22375129; PubMed Central PMCID: PMCPMC3285902.
  59. 59. Stelzhammer V, Guest PC, Rothermundt M, Sondermann C, Michael N, Schwarz E, et al. Electroconvulsive therapy exerts mainly acute molecular changes in serum of major depressive disorder patients. European neuropsychopharmacology: the journal of the European College of Neuropsychopharmacology. 2013;23(10):1199–207. pmid:MEDLINE:23183131.
  60. 60. Bilgen AE, Zincir SB, Zincir S, Ozdemir B, Ak M, Aydemir E, et al. Effects of electroconvulsive therapy on serum levels of brain-derived neurotrophic factor and nerve growth factor in treatment resistant major depression. Brain research bulletin. 2014;104:82–7. pmid:WOS:000337875300010.
  61. 61. Bumb JM, Aksay SS, Janke C, Kranaster L, Geisel O, Gass P, et al. Focus on ECT seizure quality: serum BDNF as a peripheral biomarker in depressed patients. Eur Arch Psychiatry Clin Neurosci. 2014. Epub 2014/09/19. pmid:25231834.
  62. 62. Kleimann A, Kotsiari A, Sperling W, Groschl M, Heberlein A, Kahl KG, et al. BDNF serum levels and promoter methylation of BDNF exon I, IV and VI in depressed patients receiving electroconvulsive therapy. Journal of neural transmission (Vienna, Austria: 1996). 2014. Epub 2014/11/13. pmid:25387785.
  63. 63. Marano CM, Phatak P, Vemulapalli UR, Sasan A, Nalbandyan MR, Ramanujam S, et al. Increased plasma concentration of brain-derived neurotrophic factor with electroconvulsive therapy: A pilot study in patients with major depression. Journal of Clinical Psychiatry. 2007;68(4):512–7. pmid:WOS:000245876000004.
  64. 64. Piccinni A, Del Debbio A, Medda P, Bianchi C, Roncaglia I, Veltri A, et al. Plasma Brain-Derived Neurotrophic Factor in treatment-resistant depressed patients receiving electroconvulsive therapy. European neuropsychopharmacology: the journal of the European College of Neuropsychopharmacology. 2009;19(5):349–55. Epub 2009/02/19. pmid:19223156.
  65. 65. Haghighi M, Salehi I, Erfani P, Jahangard L, Bajoghli H, Holsboer-Trachsler E, et al. Additional ECT increases BDNF-levels in patients suffering from major depressive disorders compared to patients treated with citalopram only. J Psychiatr Res. 2013;47(7):908–15. pmid:WOS:000319549300008.
  66. 66. Lin CH, Chen MC, Lee WK, Chen CC, Huang CH, Lane HY. Electroconvulsive therapy improves clinical manifestation with plasma BDNF levels unchanged in treatment-resistant depression patients. Neuropsychobiology. 2013;68(2):110–5. Epub 2013/07/25. pmid:23881232.
  67. 67. Metsis M, Timmusk T, Arenas E, Persson H. Differential usage of multiple brain-derived neurotrophic factor promoters in the rat brain following neuronal activation. Proc Natl Acad Sci U S A. 1993;90(19):8802–6. Epub 1993/10/01. pmid:8415610; PubMed Central PMCID: PMCPmc47448.
  68. 68. Khundakar AA, Zetterstrom TSC. Biphasic change in BDNF gene expression following antidepressant drug treatment explained by differential transcript regulation. Brain Res. 2006;1106:12–20. pmid:WOS:000240510900002.
  69. 69. Baj G, Leone E, Chao MV, Tongiorgi E. Spatial segregation of BDNF transcripts enables BDNF to differentially shape distinct dendritic compartments. Proc Natl Acad Sci U S A. 2011;108(40):16813–8. Epub 2011/09/22. pmid:21933955; PubMed Central PMCID: PMCPmc3189043.
  70. 70. Eriksson PS, Perfilieva E, Bjork-Eriksson T, Alborn AM, Nordborg C, Peterson DA, et al. Neurogenesis in the adult human hippocampus. Nat Med. 1998;4(11):1313–7. Epub 1998/11/11. pmid:9809557.
  71. 71. Rotheneichner P, Lange S, O'Sullivan A, Marschallinger J, Zaunmair P, Geretsegger C, et al. Hippocampal neurogenesis and antidepressive therapy: shocking relations. 2014;2014:723915. pmid:24967107.
  72. 72. Mellios N, Huang HS, Grigorenko A, Rogaev E, Akbarian S. A set of differentially expressed miRNAs, including miR-30a-5p, act as post-transcriptional inhibitors of BDNF in prefrontal cortex. Hum Mol Genet. 2008;17(19):3030–42. Epub 2008/07/18. pmid:18632683; PubMed Central PMCID: PMCPmc2722882.
  73. 73. Caputo V, Sinibaldi L, Fiorentino A, Parisi C, Catalanotto C, Pasini A, et al. Brain derived neurotrophic factor (BDNF) expression is regulated by microRNAs miR-26a and miR-26b allele-specific binding. PloS one. 2011;6(12):e28656. Epub 2011/12/24. pmid:22194877; PubMed Central PMCID: PMCPmc3237476.
  74. 74. Tongiorgi E. Activity-dependent expression of brain-derived neurotrophic factor in dendrites: facts and open questions. Neurosci Res. 2008;61(4):335–46. Epub 2008/06/14. pmid:18550187.
  75. 75. Madsen TM, Treschow A, Bengzon J, Bolwig TG, Lindvall O, Tingstrom A. Increased neurogenesis in a model of electroconvulsive therapy. Biol Psychiatry. 2000;47(12):1043–9. Epub 2000/06/23. pmid:10862803.
  76. 76. Karege F, Schwald M, Cisse M. Postnatal developmental profile of brain-derived neurotrophic factor in rat brain and platelets. Neuroscience letters. 2002;328(3):261–4. Epub 2002/07/31. pmid:12147321.
  77. 77. Elfving B, Plougmann PH, Muller HK, Mathe AA, Rosenberg R, Wegener G. Inverse correlation of brain and blood BDNF levels in a genetic rat model of depression. International Journal of Neuropsychopharmacology. 2010;13(5):563–72. pmid:WOS:000278785600002.
  78. 78. Bejot Y, Mossiat C, Giroud M, Prigent-Tessier A, Marie C. Circulating and brain BDNF levels in stroke rats. Relevance to clinical studies. PloS one. 2011;6(12):e29405. Epub 2011/12/24. pmid:22195050; PubMed Central PMCID: PMCPmc3241711.
  79. 79. Laske C, Stransky E, Leyhe T, Eschweiler GW, Maetzler W, Wittorf A, et al. BDNF serum and CSF concentrations in Alzheimer's disease, normal pressure hydrocephalus and healthy controls. J Psychiatr Res. 2007;41(5):387–94. Epub 2006/03/24. pmid:16554070.
  80. 80. Molendijk ML, Spinhoven P, Polak M, Bus BA, Penninx BW, Elzinga BM. Serum BDNF concentrations as peripheral manifestations of depression: evidence from a systematic review and meta-analyses on 179 associations (N = 9484). Molecular psychiatry. 2014;19(7):791–800. Epub 2013/08/21. pmid:23958957.
  81. 81. Molendijk ML, Haffmans JP, Bus BA, Spinhoven P, Penninx BW, Prickaerts J, et al. Serum BDNF concentrations show strong seasonal variation and correlations with the amount of ambient sunlight. PloS one. 2012;7(11):e48046. Epub 2012/11/08. pmid:23133609; PubMed Central PMCID: PMCPMC3487856.
  82. 82. Dukart J, Regen F, Kherif F, Colla M, Bajbouj M, Heuser I, et al. Electroconvulsive therapy-induced brain plasticity determines therapeutic outcome in mood disorders. Proc Natl Acad Sci U S A. 2014;111(3):1156–61. Epub 2014/01/01. pmid:24379394; PubMed Central PMCID: PMCPmc3903198.
  83. 83. Schloesser RJ, Orvoen S, Jimenez DV, Hardy NF, Maynard KR, Sukumar M, et al. Antidepressant-like Effects of Electroconvulsive Seizures Require Adult Neurogenesis in a Neuroendocrine Model of Depression. Brain stimulation. 2015. Epub 2015/07/04. pmid:26138027.
  84. 84. Sairanen M, Lucas G, Ernfors P, Castren M, Castren E. Brain-derived neurotrophic factor and antidepressant drugs have different but coordinated effects on neuronal turnover, proliferation, and survival in the adult dentate gyrus. Journal of Neuroscience. 2005;25(5):1089–94. pmid:WOS:000226750600006.
  85. 85. Egan MF, Kojima M, Callicott JH, Goldberg TE, Kolachana BS, Bertolino A, et al. The BDNF val66met polymorphism affects activity-dependent secretion of BDNF and human memory and hippocampal function. Cell. 2003;112(2):257–69. pmid:WOS:000181191600014.
  86. 86. Molendijk ML, Bus BAA, Spinhoven P, Kaimatzoglou A, Voshaar RCO, Penninx BWJH, et al. A Systematic Review and Meta-Analysis on the Association Between BDNF val(66)met and Hippocampal Volume-A Genuine Effect or a Winners Curse? American Journal of Medical Genetics Part B-Neuropsychiatric Genetics. 2012;159B(6):731–40. pmid:WOS:000308492000012.
  87. 87. Molendijk ML, van Tol MJ, Penninx BWJH, van der Wee NJA, Aleman A, Veltman DJ, et al. BDNF val(66)met affects hippocampal volume and emotion-related hippocampal memory activity. Translational Psychiatry. 2012;2. pmid:WOS:000306217600010.