Hostname: page-component-7c8c6479df-nwzlb Total loading time: 0 Render date: 2024-03-28T13:31:16.670Z Has data issue: false hasContentIssue false

Influence of electroconvulsive therapy on white matter structure in a diffusion tensor imaging study

Published online by Cambridge University Press:  23 April 2019

Jonathan Repple
Affiliation:
Department of Psychiatry, University of Muenster, Germany
Susanne Meinert
Affiliation:
Department of Psychiatry, University of Muenster, Germany
Irene Bollettini
Affiliation:
Division of Neuroscience, Scientific Institute Ospedale San Raffaele, Milano, Italy
Dominik Grotegerd
Affiliation:
Department of Psychiatry, University of Muenster, Germany
Ronny Redlich
Affiliation:
Department of Psychiatry, University of Muenster, Germany
Dario Zaremba
Affiliation:
Department of Psychiatry, University of Muenster, Germany
Christian Bürger
Affiliation:
Department of Psychiatry, University of Muenster, Germany
Katharina Förster
Affiliation:
Department of Psychiatry, University of Muenster, Germany
Katharina Dohm
Affiliation:
Department of Psychiatry, University of Muenster, Germany
Felix Stahl
Affiliation:
Department of Psychiatry, University of Muenster, Germany
Nils Opel
Affiliation:
Department of Psychiatry, University of Muenster, Germany
Tim Hahn
Affiliation:
Department of Psychiatry, University of Muenster, Germany
Verena Enneking
Affiliation:
Department of Psychiatry, University of Muenster, Germany
Elisabeth J. Leehr
Affiliation:
Department of Psychiatry, University of Muenster, Germany
Joscha Böhnlein
Affiliation:
Department of Psychiatry, University of Muenster, Germany
Ramona Leenings
Affiliation:
Department of Psychiatry, University of Muenster, Germany
Claas Kaehler
Affiliation:
Department of Psychiatry, University of Muenster, Germany Institute of Pattern Recognition and Image Analysis, University of Muenster, Germany
Daniel Emden
Affiliation:
Department of Psychiatry, University of Muenster, Germany
Nils R. Winter
Affiliation:
Department of Psychiatry, University of Muenster, Germany
Walter Heindel
Affiliation:
Department of Clinical Radiology, University of Muenster, Germany
Harald Kugel
Affiliation:
Department of Clinical Radiology, University of Muenster, Germany
Jochen Bauer
Affiliation:
Department of Clinical Radiology, University of Muenster, Germany
Volker Arolt
Affiliation:
Department of Psychiatry, University of Muenster, Germany
Francesco Benedetti
Affiliation:
Division of Neuroscience, Scientific Institute Ospedale San Raffaele, Milano, Italy University Vita-Salute San Raffaele, Italy
Udo Dannlowski*
Affiliation:
Department of Psychiatry, University of Muenster, Germany
*
Author for correspondence: Udo Dannlowski, E-mail: dannlow@uni-muenster.de
Rights & Permissions [Opens in a new window]

Abstract

Background

Electroconvulsive therapy (ECT) is a fast-acting intervention for major depressive disorder. Previous studies indicated neurotrophic effects following ECT that might contribute to changes in white matter brain structure. We investigated the influence of ECT in a non-randomized prospective study focusing on white matter changes over time.

Methods

Twenty-nine severely depressed patients receiving ECT in addition to inpatient treatment, 69 severely depressed patients with inpatient treatment (NON-ECT) and 52 healthy controls (HC) took part in a non-randomized prospective study. Participants were scanned twice, approximately 6 weeks apart, using diffusion tensor imaging, applying tract-based spatial statistics. Additional correlational analyses were conducted in the ECT subsample to investigate the effects of seizure duration and therapeutic response.

Results

Mean diffusivity (MD) increased after ECT in the right hemisphere, which was an ECT-group-specific effect. Seizure duration was associated with decreased fractional anisotropy (FA) following ECT. Longitudinal changes in ECT were not associated with therapy response. However, within the ECT group only, baseline FA was positively and MD negatively associated with post-ECT symptomatology.

Conclusion

Our data suggest that ECT changes white matter integrity, possibly reflecting increased permeability of the blood–brain barrier, resulting in disturbed communication of fibers. Further, baseline diffusion metrics were associated with therapy response. Coherent fiber structure could be a prerequisite for a generalized seizure and inhibitory brain signaling necessary to successfully inhibit increased seizure activity.

Type
Original Articles
Creative Commons
Creative Common License - CCCreative Common License - BY
This is an Open Access article, distributed under the terms of the Creative Commons Attribution licence (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted re-use, distribution, and reproduction in any medium, provided the original work is properly cited.
Copyright
Copyright © The Author(s), 2019. Published by Cambridge University Press

Introduction

Major depressive disorder is a common, recurrent and costly disorder, with a life-time prevalence of roughly 16.6% (Kessler et al., Reference Kessler, Berglund, Demler, Jin, Merikangas and Walters2005). Further, 15–33% of patients are classified as treatment resistant (Berlim et al., Reference Berlim, Fleck and Turecki2008), defined as two failed attempts to achieve significant clinical improvement with pharmacologically different antidepressants administered in adequate dose, duration and with compliance. For treatment-resistant patients, electroconvulsive therapy (ECT) is a valuable alternative as 60–80% improve and 50–60% achieve remission (Weiner and Reti, Reference Weiner and Reti2017). Understanding which patient might benefit from ECT independent of previous failed treatment attempts might therefore improve patient care.

ECT is a fast-acting intervention (Ottosson and Odeberg, Reference Ottosson and Odeberg2012) with high response rates (Fink, Reference Fink2014). In clinical practice, ECT response is linked to seizure quality parameters (Shah et al., Reference Shah, Wadoo and Latoo2013; Minelli et al., Reference Minelli, Abate, Zampieri, Gainelli, Trabucchi, Segala, Sartori, Gennarelli, Conca and Bortolomasi2016). For instance, in current clinical application, ECT treatments with short seizure lengths below 25 s are considered insufficient.

The underlying mechanisms explaining the antidepressant effect of ECT remain uncertain (Hoy and Fitzgerald, Reference Hoy and Fitzgerald2010). On the one hand, ECT seems to normalize an immune system dissemblance and inflammatory processes associated with depression (Yrondi et al., Reference Yrondi, Sporer, Péran, Schmitt, Arbus and Sauvaget2017). On the other hand, studies point to neurotrophic effects such as enhanced neurogenesis (Madsen et al., Reference Madsen, Treschow, Bengzon, Bolwig, Lindvall and Tingström2000), angiogenesis (Hellsten et al., Reference Hellsten, West, Arvidsson, Ekstrand, Jansson, Wennström and Tingström2005), gliogenesis (Wennström et al., Reference Wennström, Hellsten, Ekstrand, Lindgren and Tingström2006), synaptogenesis and heightened axonal tropism (Nickl-Jockschat et al., Reference Nickl-Jockschat, Palomero Gallagher, Kumar, Hoffstaedter, Brügmann, Habel, Eickhoff and Grözinger2016) after ECT. This neurotrophic theory is supported both by animal (Malberg et al., Reference Malberg, Eisch, Nestler and Duman2000; Kondratyev et al., Reference Kondratyev, Ved and Gale2002) and human studies (Bumb et al., Reference Bumb, Aksay, Janke, Kranaster, Geisel, Gass, Hellweg and Sartorius2015; Redlich et al., Reference Redlich, Opel, Grotegerd, Dohm, Zaremba, Bürger, Münker, Mühlmann, Wahl, Heindel, Arolt, Alferink, Zwanzger, Zavorotnyy, Kugel and Dannlowski2016). These changes might influence white matter structure to an extent that could be measured by diffusion tensor imaging (DTI). This magnetic resonance imaging technique measures diffusion of water to quantify white matter integrity.

Nonetheless, only three DTI studies investigated ECT effects yielding inconsistent results. In a longitudinal study with late-life depressed patients, mean fractional anisotropy (FA), a measure often associated with fiber integrity, increased in selected regions of interest in frontal white matter in eight patients when studying at 1.5 Tesla (T) (Nobuhara et al., Reference Nobuhara, Okugawa, Minami, Takase, Yoshida, Yagyu, Tajika, Sugimoto, Tamagaki, Ikeda, Sawada and Kinoshita2004). Lyden et al. (Reference Lyden, Espinoza, Pirnia, Clark, Joshi, Leaver, Woods and Narr2014) showed that FA increased after ECT in the anterior cingulum, forceps minor, and left superior longitudinal fasciculus when studying at 3 T voxel-based whole-brain white matter maps in 21 patients. This increase of FA was positively associated with treatment response. However, Nickl-Jockschat et al. (Reference Nickl-Jockschat, Palomero Gallagher, Kumar, Hoffstaedter, Brügmann, Habel, Eickhoff and Grözinger2016) did not detect any white matter alterations associated with ECT in 20 patients when studying at 3 T with tract-based spatial statistics (TBSS), which focuses on the most compact white matter skeleton. These inconsistent results could be due to small sample sizes, to differences in study populations, and to the major methodological differences between these studies. Furthermore, none considered seizure quality parameters or included patient control groups.

Previous studies have tried to predict ECT response using rating scales, symptom clusters, or clinical characteristics such as depressive episode length and history of medication failure (Hickie et al., Reference Hickie, Mason, Parker and Brodaty1996; De Vreede et al., Reference De Vreede, Burger and van Vliet2005; Haq et al., Reference Haq, Sitzmann, Goldman, Maixner and Mickey2015). In neurobiological studies, larger amygdala (ten Doesschate et al., Reference ten Doesschate, van Eijndhoven, Tendolkar, van Wingen and van Waarde2014), smaller inferior frontal gyrus gray matter (Oudega et al., Reference Oudega, van Exel, Stek, Wattjes, van der Flier, Comijs, Dols, Scheltens, Barkhof, Eikelenboom and van den Heuvel2014), and larger anterior cingulate cortex volume (Redlich et al., Reference Redlich, Opel, Grotegerd, Dohm, Zaremba, Bürger, Münker, Mühlmann, Wahl, Heindel, Arolt, Alferink, Zwanzger, Zavorotnyy, Kugel and Dannlowski2016) were associated with treatment response. However, no study used DTI to investigate potential predictors for treatment response.

The aims of this study were to investigate influences of ECT on white matter structure. We expect that these influences should differ from regular changes in healthy controls (HC) or treatment effects in patients receiving non-ECT treatment (NON-ECT) – inpatient treatment with medication and psychotherapy. Furthermore, white matter changes post-ECT should be associated with clinical response and seizure quality parameters. Lastly, in line with previous gray matter studies, we expect that baseline white matter coherence is a potential predictor for treatment response.

Methods and materials

Participants and study design

One hundred and seventy subjects – 106 subjects diagnosed with current major depressive disorder and 64 HC – participated in the present study. Thirty-five depressed patients were treated with ECT and 71 received in-patient treatment. Patients were recruited naturalistically with treatment being assigned based on clinical decisions independent from study participation. All subjects were diagnosed with the Structural Clinical Interview for DSM-IV-TR (Wittchen et al., Reference Wittchen, Zaudig and Fydrich1997) to confirm the psychiatric diagnosis or the lack thereof. For study inclusion and exclusion criteria, see online Supplementary Methods 1.

Thirteen subjects (10 HC and three ECT) were lost to follow-up. Two participants had to be excluded after clinical assessment, due to high self-reported depressive symptoms in the absence of a major depressive episode (one HC), or a bipolar disorder diagnosis (one NON-ECT). In the process of image quality control (see ‘Methods’ section below), six additional subjects (one HC, three ECT, one NON-ECT) had to be excluded. Therefore, the final sample comprised 29 ECT, 69 NON-ECT, and 52 HC included in all further analyses.

We conducted a non-randomized prospective study. The three groups (ECT, NON-ECT, HC) were investigated at two time points each (T 0, T 1). In the ECT group, patients were measured before treatment (T 0), and shortly after finishing treatment (T 1). The pre–post interval in control groups was 6–7 weeks adjusted for average ECT series length (Table 1). The average time between the last ECT and the post-MRI scan was 4.21 days (s.d. = 4.15). Brief-pulse ECT was conducted three times a week using an integrated instrument (Thymatron IV; Somatics Inc, Venice, FL, USA; number of sessions: M = 13.86, s.d. = 3.53). Energy dosage elevation was considered between ECT sessions if the primarily induced seizure activity lasted <25 s. For more details on ECT procedure and parameters, see online Supplementary Methods 2.

Table 1. Demographic and clinical characteristics of the sample

a Numbers present either absolute numbers or mean plus standard deviation.

b χ2-test (two-tailed).

c F-test (two-tailed).

d t test (two-tailed).

Concentration deficits: concentration deficit was rated by clinicians based on clinical presentation on a scale from 0 to 6 points. 0 resembles no difficulties in concentrating, while 6 was rated when participants were unable to read or converse without great initiative (48); CPZ, chlorpromazine equivalent doses; ECT, group treated with electroconvulsive therapy; HAMD, Hamilton Depression Rating Scale; HC, healthy controls; NON-ECT, group treated without ECT.

The three groups did not differ in days between scans, sex, and IQ (all ps > 0.176; see Table 1). We controlled for age in all subsequent analyses, as the three groups differed significantly (p < 0.001). As expected, the two patient groups ECT and NON-ECT had different clinical characteristics and medications. Patients treated with ECT had a more severe course of illness (e.g. more hospitalizations, see Table 1) and more antipsychotic medication (see Table 1) compared with NON-ECT. We used the Medication Load Index (MedIndex; Redlich et al., Reference Redlich, Almeida, Grotegerd, Opel, Kugel, Heindel, Arolt, Phillips and Dannlowski2014) – a composite measure of total medication load reflecting dose and number of prescriptions irrespective of active components – and furthermore chlorpromazine equivalent doses (CPZ; Gardner et al., Reference Gardner, Murphy, O'Donnell, Centorrino and Baldessarini2010) to measure medication intake. Medication – measured by CPZ and MedIndex – in both groups did not change between scans (all ps > 0.548). For more medication details and comorbidities, see online Supplementary Methods 3.

This study was approved by the ethics committee of the Medical Faculty of Muenster University and all subjects gave written informed consent prior to the examination. They received financial compensation for participation after the testing session.

DTI data acquisition

Data were acquired using a 3 T whole body MRI scanner (Gyroscan Intera, Philips Medical Systems, Best, the Netherlands), as reported earlier (Repple et al., Reference Repple, Meinert, Grotegerd, Kugel, Redlich, Dohm, Zaremba, Opel, Buerger, Förster, Nick, Arolt, Heindel, Deppe and Dannlowski2017). The DTI data were acquired in 36 axial slices, 3.6 mm thick with no gap (acquired matrix 128 × 128), resulting in a voxel size of 1.8 mm × 1.8 mm × 3.6 mm. The echo time was 95 ms and the repetition time was 9473 ms. A b-value of 1000 s/mm2 was used for 20 diffusion-weighted images, with isotropic gradient directions plus one non-diffusion-weighted (b 0 = 0) image. In sum, 21 images per slice were used for diffusion-tensor estimation. The total data acquisition time was approximately 8 min per subject. During the experiment, subjects lay supine in the MRI scanner with their head position being stabilized.

Image processing

Preprocessing and analysis were performed with the FSL FMRIB Software Library v10.0 [http://fsl.fmrib.ox.ac.uk/fsl/fslwiki/, FMRIB, Oxford Center for Functional MRI of the Brain, University of Oxford, Department of Clinical Neurology, John Radcliffe Hospital, Oxford, UK (Jenkinson et al., Reference Jenkinson, Beckmann, Behrens, Woolrich and Smith2012)]. For each subject, diffusion-weighted images were corrected for motion and eddy-current distortions using the eddy-correct utility in FSL with b 0 image as reference for alignment. The reference images underwent automated skull stripping before diffusion tensor estimation using the FSL FDT (Bihan, Reference Bihan2003). This yielded an estimation of FA, mean diffusivity (MD), radial diffusivity (RD), and axial diffusivity (AD) for each voxel.

To ensure data quality, all raw DTI and FA images were visually inspected. If the estimated mean displacement provided by the eddy-correct log file was larger than three times the standard deviation of all subject's mean displacement, subjects were excluded (see above). Outliers based on FA maps were detected using the homogeneity of covariance test (>2 standard deviations) provided by VBM8-toolbox (http://dbm.neuro.unijena.de/vbm).

Tract-based spatial statistics

TBSS was performed to reduce partial volume effects and registration misalignments (Smith et al., Reference Smith, Jenkinson, Johansen-Berg, Rueckert, Nichols, Mackay, Watkins, Ciccarelli, Cader, Matthews and Behrens2006). The FA images were registered to the FMRIB58 FA template and averaged to create a mean FA image. A white matter skeleton was created with an FA threshold of 0.2 and overlaid onto each subject's registered FA image. For each subject, the maximum weighted for distance FA value orthogonal to the skeleton was moved to skeleton space for group-level comparisons. The same registration was performed on MD, RD, and AD values. Factorial models for repeated measures in more than two groups are not yet implemented in FSL (Winkler et al., Reference Winkler, Ridgway, Webster, Smith and Nichols2014). Therefore, to investigate changes over time between groups, we calculated difference images subtracting pre- from post-voxel values (Ganzola et al., Reference Ganzola, Nickson, Bastin, Giles, Macdonald, Sussmann, McIntosh, Whalley and Duchesne2017). To test for statistical significance, we used the non-parametric permutation testing implemented in FSL's randomize (Winkler et al., Reference Winkler, Ridgway, Webster, Smith and Nichols2014) with 5000 permutations. Threshold-Free Cluster Enhancement (TFCE) (Smith and Nichols, Reference Smith and Nichols2009) was used to correct for multiple comparisons using the default values provided by the --T2 option optimized for TBSS. TFCE is a method for thresholding that allows richer and more interpretable outputs (Smith and Nichols, Reference Smith and Nichols2009). TFCE can be seen as a generalization of the cluster mass statistics (Bullmore et al., Reference Bullmore, Suckling, Overmeyer, Rabe-Hesketh, Taylor and Brammer1999), and uses spatial neighborhood information in a non-linear image processing to increase sensitivity and boost the height of spatially distributed signals without changing the location of their maxima. Voxel-wise levels of significance, corrected for multiple comparisons, are then calculated with standard permutation testing by building up the null distribution (across permutation of the input data) of the maximum (across voxels) TFCE scores, and then using the 95th percentile of the null distribution to threshold signals at corrected p < 0.05. This allows to estimate cluster sizes corrected for the family-wise error (FWE; p < 0.05, 5000 permutations). MNI coordinates for peak voxel and cluster sizes were derived with FSL Cluster and the corresponding white matter tract retrieved from the ICBM-DTI-81 white matter atlas (Mori et al., Reference Mori, Wakana, Nagae-Poetscher and van Zijl2006; Pekar et al., Reference Pekar, van Zijl, Reich, Zhang, Calabresi, Wakana, Mori, Li, Hua and Jiang2007).

Analysis

Statistical analyses on the demographic data were performed using IBM SPSS Statistics 25 (SPSS Inc., Chicago, IL, USA). We assessed treatment efficacy in a repeated-measures analysis of covariance (ANCOVA) in SPSS with the dependent variable sum of Hamilton Depression Scale (HAMD) (Hamilton, Reference Hamilton1960), a measure of depressive symptom severity, for each time point, the within-subject factor time (T 0 v. T 1), the between-subject factor treatment (ECT v. NON-ECT), and the nuisance variables age, sex, and days between scans.

  1. 1. Longitudinal changes in diffusion metrics (FA, MD, RD, AD) in the ECT sample were established using paired t tests within FSL.

  2. 2. For the investigation of a group by time interaction, we performed two analyses: Permutation of repeated measures, three-group ANCOVAs via ‘FSL's randomize’ is not yet implemented. Therefore, we first extracted a single mean MD value for each subject from the result mask of the paired t test (ECT T 1 v. ECT T 0). With these extracted values, we performed a proper repeated-measures ANCOVA within SPSS with group (ECT v. NON-ECT v. HC) as between-subject factor and time (T 0 v. T 1) as a within-subject factor. Age, sex, and total intracranial volume (TIV) were included as covariates. Post-hoc t tests and ANCOVAs were performed within SPSS to further investigate group differences, especially comparing the two patient groups while controlling for clinical variables (illness severity, measured by number of hospitalization and number of depressive episodes) and medication strategies (measured by MedIndex and CPZ). Second, to compare whole-brain changes in diffusion metrics (ΔFA, ΔMD, ΔAD, ΔRD) over time between groups, difference images were analyzed using t tests comparing ECT to both control groups, NON-ECT and HC, correcting for age, sex, time between scans, and TIV as recommended by FSL (https://fsl.fmrib.ox.ac.uk/fsl/fslwiki/GLM).

  3. 3. Next, to test the influence of seizure properties on white matter, difference images (ΔFA, ΔMD, ΔAD, ΔRD) were correlated with mean seizure duration, measured by seizure EEG activity in seconds. Furthermore, for exploratory analyses with additional ECT parameters, we performed correlational analyses with extracted Delta MD values (from the results mask ECT T 1 v. T 0) and ECT parameters (online Supplementary Methods 2).

  4. 4. To assess the associations between white matter integrity and treatment response in both patient groups, we conducted exploratory correlational analyses. To find associations of treatment response, baseline DTI images and changes in DTI metrics were correlated with changes in HAMD (ΔHAMD = HAMDT0 – HAMDT1), as a measure of treatment efficacy (Hamilton, Reference Hamilton1960). We repeated these analyses within the results mask from the longitudinal results (ECT T 1 v. ECT T 0), to check for anatomical overlap of these two analyses.

For all analyses with T 0 images, age, sex, and TIV were included as nuisance variables. For all analyses with difference images (ΔFA, ΔMD, ΔAD, ΔRD),, we used age, sex, TIV and days between scans as nuisance variables.

Results

Longitudinal effects

We investigated depressive symptom severity comparing HAMD values over time (T 0, T 1). We found a significant main effect of time [F (1,92) = 6.66, p = 0.011], as both treatments reduced depression severity (Table 1). We found neither a main effect of treatment (ECT v. NON-ECT) nor a treatment × time interaction (all p > 0.478) with both treatments being equally effective for the respective patient cohort.

ECT patients had significantly higher MD at T 1 compared with T 0 in several right-sided white matter tracts including the uncinate fasciculus, the posterior limb of internal capsule, the inferior longitudinal fasciculus, and inferior fronto-occipital fasciculus (p FWE = 0.018, k = 2008, x = 31, y = 0, z = −32; Figure 1, online Supplementary Results 1). There was a trend of increased AD (p FWE = 0.056) and RD (p FWE = 0.055), while no effect was detected for FA (p FWE = 0.93). No decreases of parameters were observed (all p FWE > 0.31).

Fig. 1. Longitudinal increase of MD in ECT sample. Axial slices with corresponding y-axis values (MNI) are presented. Red areas represent voxels, where a significant increase of mean diffusivity was detected after ECT in the ECT sample (p FWE < 0.05). MD, mean diffusivity.

Repeated-measures ANCOVA

A repeated-measures ANCOVA with the extracted MD values (from the significant cluster in the ECT group time effect analysis; T 0 v. T 1) revealed a significant group × time interaction [F (2,147) = 5.60; p = 0.005]. Post-hoc t tests showed that the increase in MD is significant in the ECT group [MDpre = 0.00072, MDpost = 0.00074, t(28) = 4.67, p < 0.001] and that this increase (ΔMD) is significantly higher in the ECT group compared both with NON-ECT [t(96) = 3.2, p = 0.002] and to HC [t(79) = 2.89, p = 0.005]. For a bar graph, see Fig. 2.

Fig. 2. Longitudinal effects of mean diffusivity in all groups. Bar graph displaying mean MD values (extracted from the results mask of the ECT group pre- v. post-analysis) for both time points [pre (T 0) = before first ECT treatment, post (T 1) = after last ECT treatment]. MD, mean diffusivity; ECT, group treated with electroconvulsive therapy; NON-ECT, group treated without ECT; HC, healthy controls. *Significant post-hoc increase in MD in ECT group [MDpre = 0.00072, MDpost = 0.00074, t(28) = 4.67, p < 0.001].

Furthermore, when comparing patient groups, ΔMD was significantly higher in the ECT group compared with NON-ECT even after correcting for baseline CPZ scores [F (1,92) = 17.45, p < 0.001], MedIndex [F (1,92) = 19.22, p < 0.001], ΔCPZ scores [F (1,92) = 34.45, p < 0.001], ΔMedIndex scores [F (1,92) = 33.86, p < 0.001], number of depressive episodes [F (1,92) = 34.49, p < 0.001], or number of hospitalizations [F (1,92) = 30.63, p < 0.001] in an ANCOVA already including age, sex, and TIV as covariates.

Whole-brain comparison of ΔDTI images

ΔAD was significantly higher in ECT patients compared with NON-ECT (p FWE = 0.037, k = 1251, x = −40, y = −3, z = 23) predominantly in the left anterior corona radiata (online Supplementary Results 2). There was no significant difference between ECT and NON-ECT in ΔFA (p FWE = 0.68), ΔMD (p FWE = 0.11), or ΔRD (p FWE = 0.16). ΔAD, ΔMD, ΔRD, ΔFA in ECT did not differ significantly compared with HC (all p FWE > 0.485).

Association with ECT parameters

A significant negative correlation with seizure duration in ΔFA was present (p FWE = 0.039, k = 3478, x = 7, y = 29, z = 8); longer seizure durations were associated with decreases in FA over time mostly in the corpus callosum, the corona radiata, and the internal capsule (see online Supplementary Results 3) in FSL. MD, RD, AD were not significantly associated with seizure duration (all p FWE > 0.11).

In further exploratory analyses within the ECT sample, we did not find any correlation of ΔMD with other ECT parameters such as the average electrical charge (p = 0.50), maximum electrical charge (p = 0.21), the difference score (charge last ECT – charge first ECT; p = 0.50), total number of ECT treatments (p = 0.93), days between last ECT and T 1-MRI scan (p = 0.36), and total seizure duration (across all ECT treatments, p = 0.12) in SPSS.

Association of therapy response

We found a significant positive correlation of ΔHAMD with FA (p FWE = 0.044, k = 4404, x = 36, y = −38, z = 12) and negative correlations with MD (p FWE = 0.045, k = 6107, x = 36, y = −38, z = 12) and RD (p FWE = 0.047, k = 2541, x = 33, y = −35, z = 13) at T 0 in the ECT subsample. Thus, higher FA and lower MD and RD values at baseline were positively associated with higher treatment response. The effect was mostly present in posterior fiber tracts including the splenium of corpus callosum, internal capsule, and corona radiata (Fig. 3, online Supplementary Results 4). No association was present for AD (all p > 0.152). For an additional analysis of baseline DTI measures with therapy response within the results mask of the ECT T 1 v. T 0 analyses, we again could show a positive FA and negative RD and MD correlation with ΔHAMD, for more information please see online Supplementary Material 5.

Fig. 3. Correlation of baseline (a) FA and (b) MD maps with clinical response. Top: Positive correlation FA; bottom: negative correlation MD. On the left axial slices with corresponding y-axis values (MNI) are presented. Green (FA)/red (MD) areas represent voxels, where a significant association between baseline FA/MD and ΔHAMD was found (p FWE < 0.05). Scatterplot on the right shows the association of ΔHAMD and extracted mean baseline FA/MD values from all significant voxels of corresponding TBSS analyses. FA, fractional anisotropy; MD, mean diffusivity; HAMD, sum score of the Hamilton depression scale; ΔHAMD, Difference Score (HAMDT0 – HAMDT1), high positive score reflects a good clinical response.

We did not observe an association of baseline DTI markers with treatment response in the NON-ECT group (all p FWE > 0.13). Furthermore, we did not observe an association of ΔDTI metrics with ΔHAMD in the ECT (all p FWE > 0.30) or NON-ECT group (all p FWE > 0.22).

Neither age, sex, time since first symptoms, time in weeks of current depressive episode, number of depressive episodes, number of hospitalizations, total education years, Medication Index at T 0 nor CPZ score at T 0 (all p > 0.36) were associated with ΔHAMD in the ECT subsample.

Discussion

Longitudinal changes

Within-group analyses revealed a widespread right-sided increase of MD in the ECT group after treatment compared with baseline. As most patients were stimulated unilaterally on the right side, this could underlie the laterality of the MD increase. This finding is in line with several ECT-induced gray matter changes, which are most frequently reported in the right hemisphere (Yrondi et al., Reference Yrondi, Sporer, Péran, Schmitt, Arbus and Sauvaget2017). Further, the MD increase in the ECT group was significantly greater compared with the NON-ECT group. Even more, the MD increase in the ECT group was significantly higher compared with the NON-ECT patient group when correcting for all indices of disease severity and medication, supporting the idea of ECT treatment-specific white matter alterations. The specificity of these findings is further supported by ECT parameter correlations: mean seizure duration was associated with FA decrease over time in several major longitudinal and association fibers pointing toward a negative effect of longer seizure activity on white matter integrity. On the other hand, the longitudinal changes in fiber structure were not associated with therapy response. While the increase in MD seems to be specific for ECT, it does not seem to be involved in its antidepressant effect. This is in line with the results in gray matter showing repeatedly that changes in volume post-ECT are not associated with psychopathology (Redlich et al., Reference Redlich, Opel, Grotegerd, Dohm, Zaremba, Bürger, Münker, Mühlmann, Wahl, Heindel, Arolt, Alferink, Zwanzger, Zavorotnyy, Kugel and Dannlowski2016; Oltedal et al., Reference Oltedal, Narr, Abbott, Anand, Argyelan, Bartsch, Dannlowski, Dols, van Eijndhoven, Emsell, Erchinger, Espinoza, Hahn, Hanson, Hellemann, Jorgensen, Kessler, Oudega, Paulson, Redlich, Sienaert, Stek, Tendolkar, Vandenbulcke, Oedegaard and Dale2018; Sartorius et al., Reference Sartorius, Demirakca, Böhringer, Clemm von Hohenberg, Aksay, Bumb, Kranaster, Nickl-Jockschat, Grözinger, Thomann, Wolf, Zwanzger, Dannlowski, Redlich, Zavorotnyy, Zöllner, Methfessel, Besse, Zilles and Ende2019).

A possible although speculative explanation for the MD increase could be a mild ECT-induced increase in water concentration due to an increased permeability of the blood–brain barrier. Preclinical and clinical evidence show that ECT may result in a small, transient breach in the blood–brain barrier [for review see Andrade and Bolwig (Reference Andrade and Bolwig2014)]. These white matter fiber impairments may be of clinical relevance in ECT, but future studies with additional neuropsychological testing and repeated measures during follow-up are needed to shed more light on the meaning and persistence of these alterations.

These results are in contrast to the study of Lyden et al. (Reference Lyden, Espinoza, Pirnia, Clark, Joshi, Leaver, Woods and Narr2014), who could demonstrate ECT-induced increases of FA that were associated with treatment response. However, their patient sample was younger, was tapered off of psychiatric medication, stimulus dosing was established using the titration method, and no increase of stimulation dose in the course of the ECT series was reported. Therefore, it is very likely that our sample received higher energy doses, which are known to lead to a higher antidepressive response (UK ECT Review Group, 2003), but also higher temporary cognitive impairments (Semkovska and McLoughlin, Reference Semkovska and McLoughlin2010; Tor et al., Reference Tor, Bautovich, Wang, Martin, Harvey and Loo2015).

Correlates of therapy response

Coherent fiber structure at baseline – higher FA and lower MD and RD – was positively associated with therapeutic effect (higher decrease in Hamilton Depression score) within the ECT group. It is hypothesized that the brain's ability to suppress a generalized seizure is therapeutic rather than the seizure itself (Folkerts, Reference Folkerts1996; Sackeim, Reference Sackeim1999; Kranaster et al., Reference Kranaster, Plum, Hoyer, Sartorius and Ullrich2013). Therefore, brain signaling necessary to successfully inhibit increased seizure activity might rely on intact white matter integrity. Alternatively, coherent fiber structure could be a prerequisite for a generalized seizure and its antidepressant effect. Future studies need to investigate how white matter integrity serves as a prerequisite for ECT treatment success. No other clinical or sociodemographic variables at T 0 were associated with therapy response highlighting the necessity to explore neurobiological markers for ECT response.

Limitations

We did not perform cognitive screening pre- or post-ECT treatment, restricting the interpretability of our results. Therefore, future studies need to investigate cognitive impairments with standard neuropsychological testing to further elucidate our findings.

Furthermore, the use of exploratory correlational analyses is a limitation of this study. Interpretations need to be treated with caution, as they are not corrected for multiple testing. However, they offer helpful preliminary insights into further research regarding the relationship of ECT quality parameters and white matter integrity.

The ECT and NON-ECT group differed significantly in clinical characteristics such as age, medication intake, and chronicity. The recruitment of patients in a naturalistic setting, involving different indications for drug or ECT treatment, leads to different sample characteristics, which is a clear limitation of this study. However, when correcting for age, antipsychotic medication, and illness severity, the ΔMD increase in the ECT group after treatment was still significantly higher, which points toward an ECT-specific effect on MD.

We could not detect an association of white matter changes and therapy response. While the antidepressant effect might be caused by neurotrophic mechanisms on white matter induced by ECT, it could be possible that these changes might be too subtle to be detected by our DTI scans, and moreover, they might have been overshadowed by a mild ECT-induced water increase in ΔMD. Therefore, measuring patients solely directly following treatment might constitute a limiting factor. Future studies should recruit patients repeatedly at different times after treatment (e.g. during, directly after, 2 weeks after, 6 months after) to further disentangle short-term from long-term changes.

Lastly, while our sample size is the largest reported, it is still small, which might be the reason for the absence of group differences over time between ECT and HC in the whole-brain analysis. Nonetheless, to the best of our knowledge, this longitudinal ECT DTI study is the one with the largest sample size so far, the first study to have a large depressed control group, and the first that investigated ECT parameters.

Conclusion

We could show that ECT is associated with white matter alterations. Additionally, baseline FA, MD, and RD values in a large cluster were significantly associated with therapy response in the ECT sample, which could support the future quest for a clinically suitable biomarker for therapy response prediction of ECT.

Supplementary material

The supplementary material for this article can be found at https://doi.org/10.1017/S0033291719000758

Author ORCIDs

Jonathan Repple, 0000-0003-1379-9491.

Financial support

This work was funded by the German Research Foundation (DFG, grant FOR2107 DA1151/5-1 and DA1151/5-2 to UD; SFB-TRR58, Projects C09 and Z02 to UD), the Interdisciplinary Center for Clinical Research (IZKF) of the medical faculty of Münster (grant Dan3/012/17 to UD), IMF Münster RE111604 to RR und RE111722 to RR, IMF Münster RE 22 17 07 to Jonathan Repple, and the Deanery of the Medical Faculty of the University of Münster.

Conflict of interest

V. Arolt is a member of the advisory board of, or has given presentations on behalf of, the following companies: Astra-Zeneca, Janssen-Organon, Lilly, Lundbeck, Servier, Pfizer, Otsuka, and Trommsdorff. These affiliations are of no relevance to the work described in the manuscript. H. Kugel has received consultation fees from MR:comp GmbH, Testing Services for MR Safety. This cooperation has no relevance to the work that is covered in the manuscript. The other authors declare no conflict of interest. All authors have approved the final article.

Footnotes

*

This is to indicate that the authors contributed equally to the present work and should therefore both be regarded as first/senior authors.

References

Andrade, C and Bolwig, TG (2014) Electroconvulsive therapy, hypertensive surge, blood-brain barrier breach, and amnesia: exploring the evidence for a connection. Journal of ECT 30, 160164.CrossRefGoogle ScholarPubMed
Berlim, MT, Fleck, MP and Turecki, G (2008) Current trends in the assessment and somatic treatment of resistant/refractory major depression: an overview. Annals of Medicine 40, 149159.CrossRefGoogle ScholarPubMed
Bihan, DL (2003) Looking at the functional architecture of the brain with diffusion {MRI}. Nature Reviews Neuroscience 4, 469480.CrossRefGoogle ScholarPubMed
Bullmore, ET, Suckling, J, Overmeyer, S, Rabe-Hesketh, S, Taylor, E and Brammer, MJ (1999) Global, voxel, and cluster tests, by theory and permutation, for a difference between two groups of structural MR images of the brain. IEEE Transactions on Medical Imaging 18, 3242.CrossRefGoogle ScholarPubMed
Bumb, JM, Aksay, SS, Janke, C, Kranaster, L, Geisel, O, Gass, P, Hellweg, R and Sartorius, A (2015) Focus on ECT seizure quality: serum BDNF as a peripheral biomarker in depressed patients. European Archives of Psychiatry and Clinical Neuroscience 265, 227232.CrossRefGoogle ScholarPubMed
De Vreede, IM, Burger, H and van Vliet, IM (2005) Prediction of response to ECT with routinely collected data in major depression. Journal of Affective Disorders 86, 323327.CrossRefGoogle ScholarPubMed
Fink, M (2014) What was learned: studies by the consortium for research in ECT (CORE) 19972011. Acta Psychiatrica Scandinavica 129, 417426.Google Scholar
Folkerts, H (1996) The ictal electroencephalogram as a marker for the efficacy of electroconvulsive therapy. European Archives of Psychiatry and Clinical Neuroscience 246, 155164.CrossRefGoogle ScholarPubMed
Ganzola, R, Nickson, T, Bastin, ME, Giles, S, Macdonald, A, Sussmann, J, McIntosh, AM, Whalley, HC and Duchesne, S (2017) Longitudinal differences in white matter integrity in youth at high familial risk for bipolar disorder. Bipolar Disorders 19, 158167.CrossRefGoogle ScholarPubMed
Gardner, DM, Murphy, AL, O'Donnell, H, Centorrino, F and Baldessarini, RJ (2010) International consensus study of antipsychotic dosing. American Journal of Psychiatry 167, 686693.CrossRefGoogle ScholarPubMed
Hamilton, M (1960) A rating scale for depression. Journal of Neurology, Neurosurgery, and Psychiatry 23, 5662.CrossRefGoogle ScholarPubMed
Haq, AU, Sitzmann, AF, Goldman, ML, Maixner, DF and Mickey, BJ (2015) Response of depression to electroconvulsive therapy: a meta-analysis of clinical predictors. Journal of Clinical Psychiatry 76, 13741384.CrossRefGoogle ScholarPubMed
Hellsten, J, West, MJ, Arvidsson, A, Ekstrand, J, Jansson, L, Wennström, M and Tingström, A (2005) Electroconvulsive seizures induce angiogenesis in adult rat hippocampus. Biological Psychiatry 58, 871878.CrossRefGoogle ScholarPubMed
Hickie, I, Mason, C, Parker, G and Brodaty, H (1996) Prediction of ECT response: validation of a refined sign-based (CORE) system for defining melancholia. British Journal of Psychiatry 169, 6874.CrossRefGoogle ScholarPubMed
Hoy, KE and Fitzgerald, PB (2010) Brain stimulation in psychiatry and its effects on cognition. Nature Reviews Neurology 6, 267275.CrossRefGoogle ScholarPubMed
Jenkinson, M, Beckmann, C, Behrens, TE, Woolrich, MW and Smith, SM (2012) FSL. NeuroImage 62, 782790.Google ScholarPubMed
Kessler, RC, Berglund, P, Demler, O, Jin, R, Merikangas, KR and Walters, EE (2005) Lifetime prevalence and age-of-onset distributions of DSM-IV disorders in the National Comorbidity Survey Replication. Archives of General Psychiatry 62, 593.CrossRefGoogle ScholarPubMed
Kondratyev, A, Ved, R and Gale, K (2002) The effects of repeated minimal electroconvulsive shock exposure on levels of mRNA encoding fibroblast growth factor-2 and nerve growth factor in limbic regions. Neuroscience 114, 411416.CrossRefGoogle ScholarPubMed
Kranaster, L, Plum, P, Hoyer, C, Sartorius, A and Ullrich, H (2013) Burst suppression: a more valid marker of postictal central inhibition? Journal of ECT 29, 2528.Google ScholarPubMed
Lyden, H, Espinoza, RT, Pirnia, T, Clark, K, Joshi, SH, Leaver, AM, Woods, RP and Narr, KL (2014) Electroconvulsive therapy mediates neuroplasticity of white matter microstructure in major depression. Nature Publishing Group Transl Psychiatry 4, e380.CrossRefGoogle ScholarPubMed
Madsen, TM, Treschow, A, Bengzon, J, Bolwig, TG, Lindvall, O and Tingström, A (2000) Increased neurogenesis in a model of electroconvulsive therapy. Biological Psychiatry 47, 10431049.CrossRefGoogle Scholar
Malberg, JE, Eisch, AJ, Nestler, EJ and Duman, RS (2000) Chronic antidepressant treatment increases neurogenesis in adult rat hippocampus. The Journal of Neuroscience 20, 91049110.CrossRefGoogle ScholarPubMed
Minelli, A, Abate, M, Zampieri, E, Gainelli, G, Trabucchi, L, Segala, M, Sartori, R, Gennarelli, M, Conca, A and Bortolomasi, M (2016) Seizure adequacy markers and the prediction of electroconvulsive therapy response. Journal of ECT 32, 8892.CrossRefGoogle ScholarPubMed
Mori, S, Wakana, S, Nagae-Poetscher, LM and van Zijl, PCM (2005) MRI atlas of human white matter. American Journal of Neuroradiology 27, 13841385.Google Scholar
Nickl-Jockschat, T, Palomero Gallagher, N, Kumar, V, Hoffstaedter, F, Brügmann, E, Habel, U, Eickhoff, SB and Grözinger, M (2016) Are morphological changes necessary to mediate the therapeutic effects of electroconvulsive therapy? Springer Berlin Heidelberg European Archives of Psychiatry and Clinical Neuroscience 266, 261267.CrossRefGoogle ScholarPubMed
Nobuhara, K, Okugawa, G, Minami, T, Takase, K, Yoshida, T, Yagyu, T, Tajika, A, Sugimoto, T, Tamagaki, C, Ikeda, K, Sawada, S and Kinoshita, T (2004) Effects of electroconvulsive therapy on frontal white matter in late-life depression: a diffusion tensor imaging study. Neuropsychobiology 50, 4853.CrossRefGoogle ScholarPubMed
Oltedal, L, Narr, KL, Abbott, C, Anand, A, Argyelan, M, Bartsch, H, Dannlowski, U, Dols, A, van Eijndhoven, P, Emsell, L, Erchinger, VJ, Espinoza, R, Hahn, T, Hanson, LG, Hellemann, G, Jorgensen, MB, Kessler, U, Oudega, ML, Paulson, OB, Redlich, R, Sienaert, P, Stek, ML, Tendolkar, I, Vandenbulcke, M, Oedegaard, KJ and Dale, AM (2018) Volume of the human hippocampus and clinical response following electroconvulsive therapy. Biological Psychiatry 84, 574581.CrossRefGoogle ScholarPubMed
Ottosson, JO and Odeberg, H (2012) Evidence-based electroconvulsive therapy. Acta Psychiatrica Scandinavica 125, 177184.CrossRefGoogle ScholarPubMed
Oudega, ML, van Exel, E, Stek, ML, Wattjes, MP, van der Flier, WM, Comijs, HC, Dols, A, Scheltens, P, Barkhof, F, Eikelenboom, P and van den Heuvel, OA (2014) The structure of the geriatric depressed brain and response to electroconvulsive therapy. Psychiatry Research – Neuroimaging 222, 19.CrossRefGoogle ScholarPubMed
Pekar, JJ, van Zijl, PCM, Reich, DS, Zhang, J, Calabresi, PA, Wakana, S, Mori, S, Li, X, Hua, K and Jiang, H (2007) Tract probability maps in stereotaxic spaces: analyses of white matter anatomy and tract-specific quantification. NeuroImage 39, 336347.Google Scholar
Redlich, R, Almeida, JRC, Grotegerd, D, Opel, N, Kugel, H, Heindel, W, Arolt, V, Phillips, ML and Dannlowski, U (2014) Brain morphometric biomarkers distinguishing unipolar and bipolar depression: a voxel-based morphometry-pattern classification approach. JAMA Psychiatry 71, 12221230.CrossRefGoogle ScholarPubMed
Redlich, R, Opel, N, Grotegerd, D, Dohm, K, Zaremba, D, Bürger, C, Münker, S, Mühlmann, L, Wahl, P, Heindel, W, Arolt, V, Alferink, J, Zwanzger, P, Zavorotnyy, M, Kugel, H and Dannlowski, U (2016) Prediction of individual response to electroconvulsive therapy via machine learning on structural magnetic resonance imaging data. JAMA Psychiatry 73, 557564.Google ScholarPubMed
Repple, J, Meinert, S, Grotegerd, D, Kugel, H, Redlich, R, Dohm, K, Zaremba, D, Opel, N, Buerger, C, Förster, K, Nick, T, Arolt, V, Heindel, W, Deppe, M and Dannlowski, U (2017) A voxel-based diffusion tensor imaging study in unipolar and bipolar depression. Bipolar Disorders 19, 2331.CrossRefGoogle ScholarPubMed
Sackeim, HA (1999) The anticonvulsant hypothesis of the mechanisms of action of ECT: current status. The Journal of ECT 15, 526.CrossRefGoogle ScholarPubMed
Sartorius, A, Demirakca, T, Böhringer, A, Clemm von Hohenberg, C, Aksay, SS, Bumb, JM, Kranaster, L, Nickl-Jockschat, T, Grözinger, M, Thomann, PA, Wolf, RC, Zwanzger, P, Dannlowski, U, Redlich, R, Zavorotnyy, M, Zöllner, R, Methfessel, I, Besse, M, Zilles, D and Ende, G (2019) Electroconvulsive therapy induced gray matter increase is not necessarily correlated with clinical data in depressed patients. Brain Stimulation 12, 335343.CrossRefGoogle Scholar
Semkovska, M and McLoughlin, DM (2010) Objective cognitive performance associated with electroconvulsive therapy for depression: a systematic review and meta-analysis. Biological Psychiatry 68, 568577.CrossRefGoogle ScholarPubMed
Shah, AJ, Wadoo, O and Latoo, J (2013) Electroconvulsive therapy (ECT): important parameters which influence its effectiveness. British Journal of Medical Practitioners 6(4).Google Scholar
Smith, SM and Nichols, TE (2009) Threshold-free cluster enhancement: addressing problems of smoothing, threshold dependence and localisation in cluster inference. NeuroImage 44, 8398.CrossRefGoogle ScholarPubMed
Smith, SM, Jenkinson, M, Johansen-Berg, H, Rueckert, D, Nichols, TE, Mackay, CE, Watkins, KE, Ciccarelli, O, Cader, MZ, Matthews, PM and Behrens, TEJ (2006) Tract-based spatial statistics: voxelwise analysis of multi-subject diffusion data. NeuroImage 31, 14871505.CrossRefGoogle ScholarPubMed
ten Doesschate, F, van Eijndhoven, P, Tendolkar, I, van Wingen, Ga and van Waarde, JA (2014) Pre-treatment amygdala volume predicts electroconvulsive therapy response. Frontiers in Psychiatry 5, 17.CrossRefGoogle ScholarPubMed
Tor, PC, Bautovich, A, Wang, MJ, Martin, D, Harvey, SB and Loo, C (2015) A systematic review and meta-analysis of brief versus ultrabrief right unilateral electroconvulsive therapy for depression. Journal of Clinical Psychiatry 76, e1092e1098.CrossRefGoogle ScholarPubMed
UK ECT Review Group (2003) Efficacy and safety of electroconvulsive therapy in depressive disorders: a systematic review and meta-analysis. The Lancet 361, 799808.CrossRefGoogle Scholar
Weiner, RD and Reti, IM (2017) Key updates in the clinical application of electroconvulsive therapy. International Review of Psychiatry 29, 5462.CrossRefGoogle ScholarPubMed
Wennström, M, Hellsten, J, Ekstrand, J, Lindgren, H and Tingström, A (2006) Corticosterone-induced inhibition of gliogenesis in rat hippocampus is counteracted by electroconvulsive seizures. Biological Psychiatry 59, 178186.CrossRefGoogle ScholarPubMed
Winkler, AM, Ridgway, GR, Webster, MA, Smith, SM and Nichols, TE (2014) Permutation inference for the general linear model. NeuroImage 92, 381397.Google ScholarPubMed
Wittchen, H, Zaudig, M and Fydrich, T (1997) Strukturiertes Klinisches Interview für DSM-IV, Hogrefe. Germany: Göttingen.Google Scholar
Yrondi, A, Sporer, M, Péran, P, Schmitt, L, Arbus, C and Sauvaget, A (2017) Electroconvulsive therapy, depression, the immune system and inflammation: a systematic review. Brain Stimulation.Google ScholarPubMed
Figure 0

Table 1. Demographic and clinical characteristics of the sample

Figure 1

Fig. 1. Longitudinal increase of MD in ECT sample. Axial slices with corresponding y-axis values (MNI) are presented. Red areas represent voxels, where a significant increase of mean diffusivity was detected after ECT in the ECT sample (pFWE < 0.05). MD, mean diffusivity.

Figure 2

Fig. 2. Longitudinal effects of mean diffusivity in all groups. Bar graph displaying mean MD values (extracted from the results mask of the ECT group pre- v. post-analysis) for both time points [pre (T0) = before first ECT treatment, post (T1) = after last ECT treatment]. MD, mean diffusivity; ECT, group treated with electroconvulsive therapy; NON-ECT, group treated without ECT; HC, healthy controls. *Significant post-hoc increase in MD in ECT group [MDpre = 0.00072, MDpost = 0.00074, t(28) = 4.67, p < 0.001].

Figure 3

Fig. 3. Correlation of baseline (a) FA and (b) MD maps with clinical response. Top: Positive correlation FA; bottom: negative correlation MD. On the left axial slices with corresponding y-axis values (MNI) are presented. Green (FA)/red (MD) areas represent voxels, where a significant association between baseline FA/MD and ΔHAMD was found (pFWE < 0.05). Scatterplot on the right shows the association of ΔHAMD and extracted mean baseline FA/MD values from all significant voxels of corresponding TBSS analyses. FA, fractional anisotropy; MD, mean diffusivity; HAMD, sum score of the Hamilton depression scale; ΔHAMD, Difference Score (HAMDT0 – HAMDT1), high positive score reflects a good clinical response.

Supplementary material: File

Repple et al. supplementary material

Repple et al. supplementary material

Download Repple et al. supplementary material(File)
File 319.4 KB