Neurobiological mechanisms of ECT and TMS treatment in depression: study protocol of a multimodal magnetic resonance investigation

ECT is a medical treatment in which electric currents delivered through scalp electrodes are used to intentionally trigger a brief seizure. Patients are under general anesthesia and muscular blockade during treatment sessions, which are provided two or three days per week for two to four weeks. ECT has remained the most effective acute treatment for major depressive episodes [14, 15] over the past 80 years. Even for patients with treatment-resistant, severe and sometimes life-threatening depressive episodes, the remission rates of ECT are up to 75% [16]. However, the use of ECT is still debated [17, 18], and the lack of an overarching model for the mechanism of action of ECT possibly contributes to its public controversy [19]. Considering the large number of people suffering from depression, the use of ECT is variable, with reported treated person rates (TPR; number of persons treated with ECT per 10,000 resident population per year) between 0.11 and 5.1 worldwide [20], possible due to worries about side effects.

ECT is associated with cognitive side effects [21, 22]. However, the literature is divergent and inconclusive regarding the nature and development of the neuropsychological profile of cognitive side effects [23, 24]. Most of the cognitive impairments seem to be limited to the first three days after ECT [25], and some patients even report an improvement in neurocognitive function [23]. While reduced autobiographical memory consistency has been well documented [26‐28], a number of other cognitive functions has been shown to improve one month after ECT, compared with pre-ECT [24]. Although several guidelines highlight the importance of assessing cognitive function during ECT, there is no consensus regarding the optimal test battery and study design [24, 29], thus large well-designed studies investigating broad measures of cognition pre-post ECT are warranted.

There is substantial evidence for ECT-induced volumetric changes in brain gray matter areas [30, 31], which depend on electrode placement [32] and correlate with the electrical field strength [33]. Although there are conflicting results regarding the clinical relevance of these changes, some association with clinical response has been found [34, 35], and side effects may relate to hippocampal volume change [36, 37].

TMS is a form of brain stimulation in which rapidly changing magnetic fields induce electric currents underneath a ferromagnetic coil [38]. The stimulation comes in the form of pulses and depending on the frequency of these pulses, TMS can have excitatory or inhibitory effects in cortical regions [39]. TMS is a relatively new treatment and was approved by the US Food and Drug Administration for medication resistant depression in 2008. Based on consistent findings that show hypoactivation of the left dorsolateral prefrontal cortex (DLPFC) in MDD [40, 41], a high-frequency TMS protocol was developed to re-establish normal activity levels within this region [42], to reduce symptoms and potentially improve aspects of neurocognition [43] not currently improved by traditional treatments. The treatment typically requires daily TMS sessions of up to 3000 pulses for four to six weeks [44]. The efficacy of TMS treatment has been demonstrated in several meta-analyses as well as single- and multisite studies [44, 45]. Approximately, two-thirds of patients respond to the treatment and show some improvement, while the remission rate is approximately 30% [46].

Patients are recruited when referred to ECT or TMS at Haukeland University Hospital, and healthy controls are recruited in the same geographical region as the patients. Both patients and healthy controls must understand and sign the informed consent form. All study participants receive a universal gift card to 200 Norwegian kroner (NOK) for each MRI and 300 NOK for each neuropsychological assessment in the study.

ECT will be administered with right unilateral (RUL) electrode positioning [72]. On clinical indication, electrode positioning can be switched to bilateral positioning (BL). The ECT device used is a Thymatron System IV (Somatic Inc.) that provides a brief pulse, square wave, and constant current (900 mA). The initial stimulus dose will be determined by an age-based method as described in Kessler et al. 2010 [73]. The initial duration of the stimulus pulse will be set to 0.5 ms. Treatment frequency is thrice weekly until remission or until no further improvement is expected, with an upper limit of 18 sessions. Each seizure will be evaluated on the following variables: duration, δ-waves, reorientation time and improvement in depression. The stimulus parameters can be adjusted in each session, based on evaluations of previous sessions.

Treatment is delivered with a TMS stimulator (MAG & More GmbH, München, Germany) using a figure-of-eight stimulation coil (Fo8 coil). After determination of the motor threshold at the first session, (defined at the point of maximal stimulation for the right abductor pollicis brevis or other hand muscles, as visually detected, with the paddle axis oriented laterally) the coil is placed over a point 5 cm anterior to the point at which the motor threshold is obtained. Each patient will receive 5 treatment sessions per week for 6 weeks (a total of 30 sessions). Each session comprises a protocol of 1500 TMS pulses in total, delivered in 15 trains of 100 pulses per train, with 30 s intervals at a frequency of 10 Hz, at 100% of motor threshold on the treatment site over the left DLPFC. This corresponds to 9 min and 30 s stimulation in total per session.

The clinical assessment will be performed in accordance with standard clinical protocols at the treatment unit. This includes electroencephalography (EEG), ECT stimulus specific variables (e.g., pulse width, charge delivered, amplitude, frequency, pulse train duration, seizure duration and electrode placement), reorientation time, and anesthetic used for induction. Weekly clinical assessments will include the MADRS [47], Beck Inventory test (BDI) [48], Clinical Global Impression Scale (CGI) [49], Mini Mental Status (MMSE) [51] and Everyday Memory Questionary (EMQ) [50]. Psychiatric history and the cause for cessation of ECT will be recorded. Response to treatment will be defined as > 50% reduction in MADRS score, and remission as a MADRS score ≤ 10.

We will follow a comprehensive MRI protocol which is in accordance with recommendations from the Global ECT-MRI Research Collaboration (GEMRIC) and closely follows the standards of the Adolescent Brain Cognitive Development (ABCD) study [74, 75]. This includes various imaging sequences which will allow investigation of brain structure, brain function and metabolic profiles. This protocol includes a high resolution T1 and T2 structural volumes (sagittal 3D MPRAGE T1 with isotropic voxel size of 1 mm³, echo time (TE) = 3.1 ms, repetition time (TR) = 7.4 ms, flip angel (FA) = 8 deg., time of acquisition (TA) = 6:32 min, and sagittal 3D CUBE T2 with isotropic voxel size of 1 mm³, TE = 60 ms, TR = 3200 ms, echo train length = 140, TA = 4:46 min respectively), MR spectroscopy acquired from the right amygdala (20 × 20 × 20 mm single voxel point resolved spectroscopy, SV-PRESS, TE = 35 ms, TR = 1500 ms, 128 scans of 4096 samples, spectral width = 5000 Hz, TA = 3:48 min; and a semi-LASER (PROBE-sl) single voxel 20 × 20 × 20 mm, TE = 30 ms, TR = 2000 ms, VAPOR FA = 65, 64 scans, TA = 2:28), Arterial Spin Labeling (ASL) MRI (3D ASL with effective resolution 3.64 mm, TE = 10.5 ms, TR = 4888 ms, TA = 4:39 min), resting state fMRI (rs-fMRI using 2D echo planar imaging (EPI) with isotropic voxel size of 2.4 mm³, TE = 30 ms, TR = 800 ms, TA = 10 min), and multishell diffusion weighted imaging (DWI) (1.7 mm isotropic multiband-accelerated DWI with a total of 186 diffusion encoding directions measured across four non-zero b-values (6 at b = 500, 60 at b = 1000, 60 at b = 2000, 60 at b = 3000), TA 13:49 min). High Order Shim covering the whole brain will be used for the resting state magnetic resonance imaging (rsfMRI) and diffusion tensor imaging acquisitions, and EPI field maps will be generated for each. For magnetic resonance spectroscopy (MRS), we used a semi-LASER (PROBE-sl) single voxel 20 × 20 × 20 mm, TE = 30 ms, TR = 2000 ms, VAPOR flip angle = 65, 64 scans, TA = 2:28.

A test battery with standardized and normalized neuropsychological tests will be applied to investigate aspects of cognitive functioning and possible side effects of treatment, including premorbid general abilities, cognitive status, spatial navigation, language learning, verbal- visuospatial and autobiographical memory, working memory, processing speed, attention, executive functions and inhibitory control, self-perceived cognitive difficulties, and psychomotor tempo. Based on clinical observations of spatial navigation complaints after ECT, together with the structural changes in the hippocampus [32, 76], we have choosen to include specific tests of this domain as well. The complete battery has a duration of 3–4 h and will be administered by healthcare personnel, trained in the method and under supervision of a clinical neuropsychologist.

The neurocognitive test battery is aligned with the recommendations for GEMRIC [77] with the following tests: the Rey Auditory Verbal Learning Test (RAVLT) [52] where trial 1.-7. is included and delayed recall at 30 min [53]; Phonemic Fluency (F-A-S) with 60 s of naming words on each letter, Semantic Fluency (Animals + Boy’s names) with 60 s of naming words in each category, and Trail Making Test Part A and B (TMT A + B) [54] with seconds to complete and numbers of errors; Symbol Digit Modality Test (SDMT) with total numbers of correct symbols on 120 s and Digit Span Backward [55] with a total of 8 digits; Autobiographical Memory Test (AMT) [56, 78] with total numbers of specific and categorical memories retrieved within 60 min, and average seconds memory retrieval; MoCa version 7.1 [57]; National Adult Reading Test (NART) [58] and Quick Inventory of Depressive Symptomatology (16-item) (Self-Report) (QIDS-SR 16) [59]. In addition to this, the following study will include specific neurocognitive- and clinical tests: Visual Analog Mood Scale (VAMS) [60] with a scale from 0–100 performed with a 1-item question: “How depressed do you feel right now?”, which will be assessed before and after each TMS/ECT-treatment; Spatial memory navigation tests: Memory retrieval task [61, 62], Four Mountain [63] and Sea Hero [64], which is further described below; Rey—Osterrieth complex figure test (RCFT) [65] assessed with copy and delayed retrieval; D-KEFS Color Word Interference Test (CWIT) [54] included color reading, color recognition, inhibition, and inhibition-switching, allowing for a separation of improvements in processing speed and executive functions respectively [79]; Perceived Deficits Questionary – Depression (PDQ-D) 5-item [66, 67] and Electroconvulsive Cognitive Assessment (ECCA) [68].

To assess laterality effects, we adopt a dichotic listening paradigm, a well-established method to determine language specialization [69]. In dichotic listening, two different syllables are presented simultaneously to the left and right ear (e.g., “ba” to the left and “ta” to the right ear). Participants tended to report the right ear stimulus because of stronger contralateral connections with the language-specialized left hemisphere. The present study will employ a dichotic listening paradigm implemented as a smartphone app [80].

Three tasks for assessing spatial navigation will be used. Although spatial navigation is a function known to be located in the hippocampal region [81], there is a lack of studies investigating spatial navigation in relation to ECT.

The memory retrieval task [61, 62] is performed on a computer screen, where the participant is moving in a virtual grassy plane surrounded by a circular wall. Behind the wall, the scenery contains landmarks such as mountains, the sun, and clouds. Objects appear at certain spots, and the participant uses the keyboard to navigate around the plane and pick up the objects. The task is to remember the location and replace the objects back to the same location.

The Four Mountains test [63] is administered using a tablet computer. First, a picture of a four-mountain landscape shows up. After a brief delay, four similar pictures appear, one of which is the same landscape as the first one but viewed under different conditions (such as a change in viewpoint or season). The task is to choose the correct landscape matching the first shown. The test includes 15 trials with different mountain landscapes.

Sea Hero Quest is an application from the lab of Hugo Spiers, University College London [64]. This test is downloadable to a handheld device, and for the current study, we use a tablet computer. The task is to first look at a map of a sea landscape with a route to follow. The map closes, and the player, which is steering a boat, should remember the map and navigate to find the target location, as previously illustrated on the map. Another task is to remember where you came from, point in that direction, and shoot a projectile back in the direction of the right spot.

A research-grade ambulatory wristband (Empatica E4) [71] with sensors that register accelerometric data, peripheral body temperature, electrodermal activity, and photoplethysmography (PPG) [82‐84] is applied. PPG employs infrared light to gauge volumetric changes in blood circulation, making it a widely used method for estimating heart rate and obtaining pulse oximetry readings [85, 86]. Heart rate variability analysis displays the function of the autonomic nervous system, and mood disorders are linked with autonomic dysfunction [87]. Autonomic activity can furthermore be assessed by observing electrodermal activity (skin conductance), which measures sweating on the skin, a function regulated by the sympathetic branch of the autonomic nervous system [88]. Accelerometry measures motor activity, with recordings of acceleration in three-dimensional space over time. The information in motor activity is hypothesized to be an objective observation of the inner physiological state [89]. Variations in energy-levels and activation patterns provide information about mood state, as well as circadian rhythmicity, which seems disrupted in depression [90, 91]. Additionally, the peripheral body temperature also expresses circadian rhythmicity [92]. These data will be acquired over a 48-h period to assess the circadian effects of the heart, electrodermal activity, motor activity and body temperature.

Based on a previous study [93], the first hypothesis of this protocol is that baseline global cerebral blood flow (CBF) will separate responders (lower CBF) from nonresponders at a threshold of ~ 45 mL/100 g/min, and that CBF will increase with ECT. The sample size needed to detect a difference at baseline between responders and nonresponders based on the data in Fig. 1 in [93] is 14 per group (1- β = 0.8, α = 0.05). Detection of a treatment-induced change in the regional hippocampal CBF (paired t test) will require a sample of 22 patients. Our sample of 50 patients should be adequate.

Very large sample sizes are needed to boost statistical power and close knowledge gaps (e.g., confirm or refute results from small studies) and to utilize machine-learning tools for outcome prediction. Hence, team science efforts are important, and parts of the analysis will be achieved through data sharing and collaboration with consortia such as the Genomics of ECT International Consortium [94](Gen-ECT-ic) and GEMRIC [77].

In addition to the two primary hypotheses, several analyses can be performed that can shed light on the mechanisms of action suggested by the DPR hypothesis.