Comparing the Neurocognitive Effects of Right-Unilateral Ultra-Brief Pulse Electroconvulsive Therapy and Magnetic Seizure Therapy for the Treatment of Major Depressive Episode

Abstract

Background:

Magnetic seizure therapy (MST) is under investigation as a treatment for adults with major depression. Prior research suggested that MST has comparable antidepressant efficacy with electroconvulsive therapy (ECT), but with greater cognitive safety. The objective of the study was to compare the neurocognitive outcomes of patients receiving an acute course of MST with those receiving ECT for the treatment of major depressive episode.

Methods:

This was a between-subject, double-masked, randomized, multi-center clinical trial. Seventy-three participants with a severe major depressive episode were enrolled and randomly assigned to treatment with MST (N = 38) or ultra-brief pulse right unilateral ECT (N = 35). The main outcomes were change in performance from baseline to end of acute treatment on multiple neurocognitive measures.

Results:

Patients receiving MST, relative to those receiving ECT, had superior cognitive outcomes up to 72-hours post-treatment. Specifically, following MST treatment, there was significant improvement in fine motor dexterity (p=0.017) and no significant change in cognitive domains of attention, verbal fluency, executive function, and verbal learning and memory. In contrast, following treatment with ECT, patients demonstrated significantly worsened performance on measures of verbal fluency (p<0.001), executive function (p=0.038), and verbal memory retention (p<0.001). Autobiographical memory consistency significantly decreased following treatment with both ECT (p<0.001) and MST, though the magnitude of change was greater for ECT.

Conclusions:

The study findings confirm prior work and provide new evidence supporting the enhanced cognitive safety of MST relative to ECT. Future research is warranted on MST to optimize its application in individuals across the lifespan with neuropsychiatric illnesses.

Clinical Trial Registration:

ClinicalTrials.gov identifier: NCT00488748

Introduction

Major depressive disorder (MDD) is a complex neuropsychiatric illness that is often treatment-resistant(1, 2). Antidepressant strategies for MDD, particularly when treatment-resistant(3), have included transcranial magnetic stimulation (TMS)(4), ketamine(5), and electroconvulsive therapy (ECT)(6). Spanning approximately eight decades, consistent evidenced has substantiated the antidepressant benefits of ECT for adults across the lifespan with treatment-resistant MDD(7, 8). While it is used in clinical practice across the globe and continues to be studied to better understand its mechanisms of action, its use at times is precluded due to its adverse cognitive effects(9–11). Magnetic seizure therapy (MST; see Figure S1) is a non-invasive brain stimulation technology under development as a safer alternative to ECT(12, 13). MST involves the use of transcranial magnetic stimulation (TMS) technology with magnetic coils placed over targeted brain regions (e.g., prefrontal cortex) to induce a therapeutic seizure (while under anesthesia) to treat depression(13, 14). Research demonstrates that MST is feasible, medically safe, and cognitively benign(13, 15–19), and carries significant antidepressant benefits(20).

Two randomized clinical trials (RCTs) that compared ECT and 100Hz MST found that both produced similar antidepressant outcomes, but MST resulted in significantly less cognitive difficulty than ECT(21, 22). Those studies warrant replication due to small sample sizes and neurocognitive outcomes that primarily centered around memory function. Furthermore, prior work did not compare MST neurocognitive outcomes with the form of ECT that has demonstrated efficacy with the fewest cognitive adverse effects, that being ultra-brief pulse right unilateral (UB-RUL ECT) at six times the seizure threshold(5, 23–27).

In the largest RCT to date that compared the antidepressant and side effects of ECT and MST, we showed that UB-RUL ECT at six times the seizure threshold and MST completers had relatively similar antidepressant response (ECT: 62.5%, MST: 58.6%) and remission (ECT: 41.7%, MST: 44.8%) rates. These relatively similar antidepressant response (ECT: 42.1%, MST: 51.4%) and remission (ECT: 26.3%, MST: 37.1%) rates between both the ECT and MST groups were also found in the intent-to-treat analysis.(24) As opposed to ECT; however, patients treated with MST showed no significant decline in performance from baseline to end of acute treatment on a neuropsychological measure of autobiographical memory and autobiographical memory specificity. In this ancillary study of our recently published RCT, here we report the comprehensive acute neurocognitive outcomes in adults with major depressive episode treated with either ECT or MST(24). We hypothesized that compared to ECT, patients who received MST would show no significant decline in performance on cognitive measures from baseline to end of the acute treatment course.

Methods and Materials

Study Design

The methods of this ancillary study have been detailed elsewhere(24). In brief, this was a three-center, between-subject, double-masked, RCT comparing the antidepressant efficacy and neurocognitive effects of MST and ECT. This study was conducted under an Investigational Device Exemption (IDE) from the United States Food and Drug Administration (FDA) and was approved by the Institutional Review Boards (IRB) at the three study sites (New York State Psychiatric Institute, Duke University Medical Center, and UT Southwestern Medical Center). All study participants provided written informed consent following the study sites’ IRB regulations.

Randomization and Masking

Participants were randomly assigned 1:1 to MST or ECT with simple randomization by a computer-generated random number sequence that was stratified by study site. Participants and raters were masked to treatment allocation, and as previously reported, the mask was maintained for both participants and raters(24).

Participants

Seventy-three patients were enrolled after providing written informed consent (see Figure 1, Consort Diagram). Inclusion criteria included men and women between the ages of 18–90 who were referred for ECT, with a MDE in the context of MDD or bipolar disorder based on the Structured Clinical Interview for the DSM-IV-TR (SCID), and a baseline 24-item Hamilton Depression Rating Scale(28) (HDRS₂₄) total score ≥ 18. Participants were eligible for the study if they had psychotic symptoms and suicidal ideation (without attempt). Exclusion criteria included a history of neurological disorders, head trauma, contraindications to TMS, current unstable or serious medical illnesses, pregnancy/breast-feeding, history of ECT in the prior 6-months or prior failure to respond to ECT, or a Mini Mental State Examination(29) (MMSE) total raw score below 24.

Figure 1. Consort Diagram.

Consort diagram showing number of participants from enrollment phase through intent-to-treat sample analysis.

Psychotropic medications, except for lorazepam up to 3mg/day as needed, was washed out for at least 3 days prior to starting treatment. As recommended by the American Psychiatric Association Task Force Report on ECT(30), benzodiazepines were withheld 10 hours prior to treatment.

Intervention

MST was administered with a Magstim Theta device (Magstim Co., Whitland, Wales, UK) with a double-layer round coil (44-mm inner diameter, 120-mm outer diameter) positioned on the vertex. UB-RUL ECT was administered with either a spECTrum 5000Q (MECTA Corp., Lake Oswego, OR, USA; 0.3 ms pulse width; 800 mA pulse amplitude) or Thymatron System IV (Somatics, Inc., LLC, USA & Canada; 0.25 ms pulse width; 900 mA pulse amplitude) using ultra-brief pulse width, empirical dose titration, and RUL electrode placement.

Seizure threshold (ST) was titrated at the first session. For MST, stimulus frequency is fixed at 100Hz, pulse intensity is set to maximum stimulator output, and train duration of 10-seconds. For MST titration, the first step was given at 5-second train duration; if there was no seizure, the duration was increased to 10-seconds. Subsequent MST treatments were given 10-second duration, the maximum duration for the device, for a total of 1,000 pulses. For ECT, dose titration followed published recommendations(30). Subsequent ECT treatments were administered at 6×ST. Both MST and ECT were performed under the same general anesthesia regimen

Patients were prepped in the same manner (i.e., earplugs placed, skin prepped for ECT electrode placement) prior to the start of each treatment regardless of treatment randomization. Auditory stimulus of a recorded MST session was played while ECT was administered to help preserve masking. Patients received three treatments per week until they achieved remission or until there was a plateau in therapeutic response. If there was < 25% decrease in HDRS₂₄ from baseline by the 8^th treatment, treatment was discontinued, and the patient was provided routine clinical care. Completers were defined as having received ≥ 8 treatments or met remission criteria prior to the 8^th treatment.

Clinical Assessment Procedures

The primary and secondary clinical outcome measures were the HDRS₂₄ and the clinician-rated Inventory of Depressive Symptomatology(31) (IDS-C), respectively. Other clinical outcomes included the self-report IDS (IDS-SR), Clinical Global Impression - Improvement(32) (CGI-I), and Global Assessment of Functioning(33) (GAF). Clinical assessments were conducted by masked raters at baseline, the morning of each treatment, and 24–72 hours after the last treatment.

Neurocognitive Assessment Procedures

The study was designed to assess multiple neurocognitive domains including attention, autobiographical memory and autobiographical memory consistency, executive function, learning and memory, motor function, processing speed, subjective cognitive function, verbal fluency, and working memory. Specific neuropsychological instruments included: Autobiographical Memory Interview Short-Form (AMI-SF)(34), Autobiographical Memory Test (AMT; results were previously reported)(24, 35), Cognitive Failures Questionnaire (CFQ)(36), Complex Figure Test (CFT)(37, 38), Delis Kaplan Executive Function System (DKEFS) Sorting Test(39), Grooved Pegboard Test(40), Rey Auditory Verbal Learning Test (RAVLT)(41), Trail Making Test Parts A and B(42), Controlled Oral Word Association Test and Category Fluency Test(43), and the Wechsler Adult Intelligence Scale-3^rd Edition (WAIS-III) Digit Span Test(44). Most of the neuropsychological measures had an alternate form and available demographic-adjusted scores. Since not all measures had available demographic-adjusted scores, all analyses reported are in raw scores. Global cognitive function was assessed with the Mini Mental State Examination(29) (MMSE; results were previously reported)(24). Estimated premorbid intellectual ability was assessed with the Wechsler Test of Adult Reading (WTAR)(45). Except as noted above, the neurocognitive assessment results from this study have not been previously reported.

Assessments were conducted by masked raters at baseline and 24–72 hours after the last treatment. Raters were trained to criteria via in-person and virtual interactive neurocognitive assessment training sessions that followed the neuropsychological instrument and neurocognitive procedure manual.

Statistical Analysis

We report mixed-design analyses of covariance (ANCOVAs) to investigate the effects of treatment group (ECT vs. MST, between groups) and time (before and after the acute treatment course) and potential interactions, while controlling for sex, age, and education, on neurocognitive assessment performance on the intent-to-treat sample. We also examined effects of study site, which was only associated with performance on the Letter Fluency Test (see Supplemental Materials). Analyses were significant at two-tailed p<.05. Post-hoc tests, using the Bonferroni adjustment, were used to follow-up significant interactions and main effects. All confidence intervals are reported at 95%. Missing data were treated as missing and not imputed. Analyses were conducted using IBM SPSS (version 29).

Results

Baseline Demographic and Clinical Characteristics

The total intent-to-treat sample enrolled (Table 1) was 56% female, 44% male, 84% White, approximately 48.0 years old (SD=14.1, range: 19–80), and with 15.8 years of education (SD=2.8). Clinically, the majority (86.3%) of participants had a major depressive episode (MDE) in the context of MDD. The current MDE was severe (HRSD₂₄ total score = 31.0 (SD=7.1) with a length of 124.2 weeks (SD=169.1). There were no statistically significant differences in demographic and clinical characteristics between treatment groups at baseline.

Table 1.

Baseline demographic and clinical characteristics of the whole study sample and each treatment group

	All	ECT	MST	t	χ 2	p
N	73	38	35
Age, mean (SD), y	48.0 (14.1)	48.2 (12.8)	47.7 (15.6)	0.12		.90
Sex, No. female (%)	41 (56.2)	19 (50)	22 (62.9)		0.76	.38
Race*, No. Caucasian (%)	62 (84.9)	34 (89.5)	28 (80)		0.65	.42
Education, mean (SD), y	15.8 (2.8)	15.5 (2.5)	16.3 (3.2)	1.19		.24
Bipolar, No. (%)	10 (13.7)	6 (15.8)	4 (11.4)		0.04	.84
Current depressive episode, mean (SD), weeks	124.2 (169.1)	114.5 (129.4)	135.2 (208.3)	0.39		.70
Number of Psychotropic medications, mean (SD)	7.6 (4.5)	7.8 (4.5)	7.3 (4.5)	0.41		.68
Benzodiazepine (No. (%))	11 (15)	6 (16)	5 (14)
Antidepressant (No. (%))	63 (86)	32 (84)	31 (89)
Antipsychotic (No. (%))	29 (40)	13 (34)	16 (46)
Lithium Augmentation (No. (%))	18 (25)	7 (18)	11 (31)
HDRS24, mean (SD) Baseline	31 (7.1)	31.8 (7.8)	30.1 (6.4)	1.06		.29
WTAR	110.85 (8.14)	110.53 (8.14)	111.20 (8.25)	.35		.73

Note. ECT = Electroconvulsive therapy; HDRS₂₄ = 24-item Hamilton Rating Scale for Depression; MST = Magnetic seizure therapy; WTAR = Wechsler Test of Adult Reading. Due to small sample sizes in other racial/ethnicity categories, reporting is omitted to prevent identifiability.

Change in Neurocognitive Performance after ECT and MST

Following acute treatment with ECT and MST, there were differential effects across neurocognitive measures with performance on some measures worsening in the ECT treatment condition, but performance across all measures remaining relatively similar or improving in the MST treatment condition in the intent-to-treat sample (Table 2) and in the completer analysis (Table S2), which showed a comparable pattern of results. Importantly, the ANCOVAs showed that there were no significant differences between treatment groups on any neurocognitive variable at baseline.

Table 2.

Neurocognitive assessment scores between electroconvulsive therapy and magnetic seizure therapy at baseline and end of the acute treatment course.

Neurocognitive Assessment	Electroconvulsive Therapy (n=38)		Magnetic Seizure Therapy (n=35)		Interaction of Treatment and Time	Main Effect of Time	Main Effect of Treatment
	Baseline	End	Baseline	End
Attention and Processing Speed
Trail Making Test A	41.03 (18.45)	45.52 (19.24)	34.94 (16.52)	32.13 (15.35)	F1,55=5.360, p=0.024, ηp2=0.089	F1,55=1.227, p=0.273, ηp2=0.022	F1,55=3.066, p=0.086, ηp2=0.053
Grooved Pegboard Test – Dominant Hand	83.93 (23.44)	85.07 (17.77)	80.47 (25.72)	73.47 (14.26)	F1,55=4.068, p=0.049, ηp2=0.069	F1,55=.475, p=0.494, ηp2=0.009	F1,55=0.842, p=0.363, ηp2=0.015
Grooved Pegboard Test – Nondominant Hand	99.45 (34.67)	99.93 (31.46)	85.37 (20.05)	79.78 (16.28)	F1,56=4.054, p=0.049, ηp2=0.068	F1,56=0.226, p=0.636, ηp2=0.004	F1,56=3.876, p=0.054, ηp2=0.065
WAIS-3 Digit Span Forward	11.16 (2.62)	11.19 (2.54)	10.91 (2.36)	11.16 (2.17)	F1,58=.176, p=0.677, ηp2=0.003	F1,58=.321, p=0.573, ηp2=0.006	F1,58=1.482, p=0.228, ηp2=0.025
WAIS-3 Digit Span Backward	6.97 (2.87)	6.39 (2.79)	7.41 (1.85)	7.78 (2.43)	F1,58=3.969, p=0.051, ηp2=0.064	F1,58=0.336, p=0.565, ηp2=0.006	F1,58=1.176, p=0.283, ηp2=0.020
Verbal Fluency
Letter Fluency	39.06 (14.42)	31.55 (13.81)	43.00 (11.97)	46.50 (10.55)	F1,58=12.768, p<0.001, ηp2=0.180	F1,58=0.720, p=0.400, ηp2=0.012	F1,58=8.076, p=0.006, ηp2=0.122
Category Fluency	15.71 (5.18)	12.48 (4.86)	16.78 (7.10)	15.63 (5.33)	F1,58=1.304, p=0.258, ηp2=0.022	F1,58=1.517, p=0.223, ηp2=0.025	F1,58=1.497, p=0.226, ηp2=0.025
Executive Function
DKEFS Sorting Test Total Score	4.86 (1.53)	4.34 (1.63)	5.23 (1.07)	5.60 (1.22)	F1,54=6.593, p=0.013, ηp2=0.109	F1,54=0.711, p=0.403, ηp2=0.013	F1,54=3.942, p=0.052, ηp2=0.068
Trail Making Test B	104.28 (73.33)	97.56 (57.24)	68.43 (36.20)	69.00 (30.28)	F1,48=0.304, p=0.584, ηp2=0.006	F1,48=0.621, p=0.435, ηp2=0.013	F1,48=4.970, p=0.031, ηp2=0.094
Learning and Memory
AMI-SF Total Score	53.67 (9.88)	42.30 (11.54)	55.45 (4.07)	48.48 (6.62)	F1,54=4.558, p=0.037, ηp2=0.078	F1,54=0.520, p=0.474, ηp2=0.010	F1,54=2.257, p=0.139, ηp2=0.040
AMI-SF Percent Consistency Recall	--	79.60 % (16.67 %)	--	87.35 % (9.47 %)	--	--	F1,54=4.391, p=0.041, ηp2=0.075
RAVLT Trials 1–5 Free Recall Total Correct	43.35 (9.94)	38.94 (11.70)	46.53 (10.08)	46.34 (9.52)	F1,58=2.949, p=0.091, ηp2=0.048	F1,58=0.454, p=0.503, ηp2=0.008	F1,58=2.461, p=0.122, ηp2=0.041
RAVLT Immediate Recall Score	8.19 (3.31)	6.39 (3.49)	9.83 (3.34)	9.52 (3.64)	F1,55=3.420, p=0.070, ηp2=0.059	F1,55=0.611, p=0.438, ηp2=0.011	F1,55=5.875, p=0.019, ηp2=0.097
RAVLT Delayed Recall Score	7.74 (3.53)	6.42 (6.38)	9.45 (3.59)	9.16 (3.71)	F1,57=1.009, p=0.319, ηp2=0.017	F1,57=0.713, p=0.402, ηp2=0.012	F1,57=3.625, p=0.062, ηp2=0.060
RAVLT Percent Retention	70.32 (22.47)	52.42 (23.71)	78.06 (32.22)	76.03 (21.71)	F1,57=6.069, p=0.017, ηp2=0.096	F1,57=0.038, p=0.846, ηp2=0.001	F1,57=7.969, p=0.007, ηp2=0.123
RAVLT Delayed Recognition Discriminatio n Score	10.00 (4.24)	6.97 (4.32)	11.03 (3.86)	10.00 (4.62)	F1,55=3.256, p=0.077, ηp2=0.056	F1,55=0.007, p=0.934, ηp2<0.001	F1,55=3.560, p=0.064, ηp2=0.061
Subjective Cognitive Function
CFQ Total Score	50.54 (13.67)	44.82 (12.43)	51.59 (14.39)	41.93 (12.12)	F1,52=1.278, p=0.263, ηp2=0.024	F1,52=0.104, p=0.749, ηp2=0.002	F1,52=0.211, p=0.648, ηp2=0.004

Note. AMI-SF = Autobiographical Memory Interview – Short Form, CFQ = Cognitive Failures Questionnaire, DKEFS = Delis Kaplan Executive Function System, MMSE = Mini Mental Status Examination, RAVLT = Rey Auditory Verbal Learning Test, WAIS-3 = Wechsler Adult Intelligence Scale – 3^rd Edition. Descriptive statistics are presented as M (SD). For certain assessments, data are missing and treated as missing (i.e., no imputations were conducted). Thus, actual sample sizes per assessment may vary slightly. Sample sizes and analyses reflect the intent-to-treat sample. Bolded cells highlight significant tests at p<0.05.

Attention and Processing Speed

Simple visual scanning/attention and psychomotor processing speed over the acute treatment course remained similar following both ECT and MST. There was a significant interaction between treatment condition and time on the Trail Making Test Part A. At the end of the study, the MST relative to the ECT group had significantly faster completion times (p=0.013, CI=(2.38,19.16)).

Fine motor dexterity and speed remained similar following ECT, but significantly improved following MST. There was a significant interaction between treatment condition and time on the Grooved Pegboard test for both the dominant and non-dominant hands. The MST group had significant improvement from baseline to study end with both the dominant (p=0.017, CI=(−13.41, −1.40)) and non-dominant hand (p=0.024, CI=(−12.90, −0.96)). At the end of the study, the MST group also had significantly faster times than the ECT group with the dominant hand (p=0.026, CI=(−15.98, −1.05)) and non-dominant hand (p=0.012, CI=(−28.04, −3.69)).

Both simple and complex auditory attention remained similar following both ECT and MST. There was no interaction, main effect of treatment condition, or main effect of time on both the Digit Span Forward and Backward tests.

Verbal Fluency and Executive Function

Phonemic fluency significantly worsened following ECT, but remained unchanged following MST. On the Letter Fluency task, there was significant interaction such that the ECT group exhibited significant worsening from baseline to study end (p<0.001, CI=(11.52,−3.15)), while the MST group performance showed no change over time. Also, at study end, the MST group significantly outperformed the ECT group (p<0.001, CI=(7.10, 19.61)). However, semantic fluency on the Category Fluency task remained similar following both ECT and MST, as there was no significant main effect of treatment condition, time, or interaction.

Concept formation and cognitive flexibility significantly worsened following ECT but remained relatively the same following MST. There was no significant effect of time or treatment condition on the DKEFS Sorting Test, but there was a significant interaction between treatment condition and time. The ECT group experienced a significant decrease in scores from baseline to end (p=0.038, CI=(−1.03, −0.03)), while the MST group did not. At the end of the study, the MST group had significantly higher scores than the ECT group (p=0.006, CI=(0.31,1.71)).

Complex visual scanning/attention, psychomotor processing speed, and cognitive flexibility was measured with the Trail Making Test Part B. On this task, there was no effect of time or interaction, but there was a significant treatment effect, such that the MST group had significantly faster times than did the ECT group (p=.03, CI=(2.59,50.19)).

Learning and Memory and Autobiographical Memory Consistency

Verbal learning, delayed recall, and delayed recognition on the RAVLT test remained similar following both ECT and MST as there was no significant effect of treatment condition, time, or interaction. However, both immediate recall (p=0.004, CI=(−3.108, −0.64)) and percent retention (p<0.001, CI=(8.77, 29.72)) of words significantly decreased following ECT, but remained relatively similar following MST. For immediate recall of words, there was no significant effect of time or interaction, but there was an effect of treatment condition, such that the MST group recalled significantly more words than the ECT group (p=0.019, CI=(0.32,3.39)). For percent retention of words (Figure 2a), there was a significant interaction. The ECT group experienced a significant decrease from baseline to study end (p<0.001, CI=(−29.72, −8.77)). Additionally, at study end, the MST group outperformed the ECT group (p<0.001, CI=(12.64,36.75)).

Figure 2. Anterograde and Retrograde Amnesia is Clinically Meaningful for Electroconvulsive Therapy, but not Magnetic Seizure Therapy.

Figure 2a. Rey Auditory Verbal Learning Test (RAVLT) percent retention scores by treatment group. A line plot displays the change of RAVLT percent retention scores by treatment group and time point. Error bars represent standard error of the mean.

Figure 2b. Autobiographical Memory Interview-Short Form (AMI-SF) recall scores by treatment group. A line plot displays the change of AMI-SF recall scores by treatment group and time point. Error bars represent standard error of the mean.

We computed the statistics for autobiographical memory consistency, as measured by the AMI-SF, based on prior published studies that contrasted the baseline total scores and follow-up total scores(20, 21, 46) and examined the percent consistency recall score(47, 48). Autobiographical memory consistency significantly decreased following both ECT and MST, though the magnitude of change was less for MST relative to ECT (Figure 2b). There was a significant interaction between treatment condition and time. Specifically, both groups had decreased performance from baseline to end (p<0.001; ECT CI=(−14.20, −8.53), MST CI=(−9.86, −4.08)), and the MST group outperformed the ECT group at study end (p=0.044, CI=(0.15,10.45)). There was no main effect of treatment condition or time. Additionally, we calculated percent autobiographical memory consistency recall scores by taking the proportion between the total AMI-SF total score at study end and baseline and multiplying by 100(34). An ANCOVA that controlled for age, sex, and years of education showed that relative to the ECT group, the MST group had significantly better percent consistency recall (p=0.041).

Subjective Cognitive Function

Subjective cognitive function remained relatively similar across the acute treatment course in both the ECT and MST conditions. On the CFQ, there was no significant effect of time, treatment condition, or interaction between treatment condition and time.

Discussion

To our knowledge, this is the first study to compare the neurocognitive effects of acute antidepressant treatment with UB-RUL ECT and MST in adults with severe MDE. The central finding is that patients receiving MST, relative to those receiving ECT, had superior cognitive outcomes. This finding is remarkable given the selection of UB-RUL ECT, which has better cognitive outcomes than other forms of ECT(26), as the comparator condition. Following MST, with the exception on an autobiographical memory consistency measure, performance across neurocognitive measures remained relatively the same from baseline to end, and in some instances improved suggesting that MST is cognitively safe. Specifically, there was significant improvement in fine motor dexterity and stability in cognitive domains of simple and complex attention, verbal fluency, executive function, verbal learning and memory, and as we previously reported also in time to reorientation and autobiographical memory and specificity(24). In contrast, following ECT, patients demonstrated worsened performance on measures of attention, processing speed, verbal fluency, executive function, learning and memory, and autobiographical memory consistency. Since both MST and ECT produce generalized seizures, but differ substantially in the intensity and depth of the E-field induced in the brain, our findings support the hypothesis that reducing E-field exposure is an effective means of reducing the cognitive side effects of seizure therapy(7, 49, 50). However, as we previously reported, seizure duration was associated with time to orientation(24), which suggests that the seizure, in addition to the E-field, may independently or interactively impact cognitive outcomes(50, 51). Additional research is warranted to inform the independent and combined contributions of the E-field and seizure to changes in neurocognitive performance.

While ECT and MST have similar antidepressant benefits(24), they have vast differences in their neurocognitive effects. Indeed, transient anterograde amnesia for new information and retrograde amnesia for autobiographical memory recall and consistency are hallmark cognitive difficulties observed with ECT(9, 10, 46, 52), but not with MST(20, 53) (refer to Figure 2b). Anterograde amnesia has been found to onset during the acute ECT course and persist for up to one-month(54) and is moderated by patient demographic factors and ECT parameters(11). Our study findings of anterograde amnesia are consistent with relatively recent studies that also examined such effects of UB-RUL ECT(46, 47). Importantly, our findings replicate and provide new evidence supporting that MST does not produce anterograde amnesia in clinical studies(16, 20, 55) or in preclinical models which showed that performance on anterograde memory measures following MST was no different than anesthesia-alone sham(18, 56).

Retrograde amnesia is associated with ECT parameters and patient demographics, and research has found that it can persist for up to six-months and perhaps beyond(57). The findings of this study are similar to other studies that examined the autobiographical memory consistency effects of UB-RUL ECT(27, 46). However, significant change in autobiographical memory consistency in MST was most likely an effect of time rather than of the treatment itself. This finding is consistent with prior clinical studies of MST(20, 21, 53) that found the rate of change on autobiographical memory consistency was relatively similar to the change over time observed in healthy adult controls(57, 58). Moreover, AMT performance demonstrated stable autobiographical memory and autobiographical memory specificity following the acute MST course suggesting that MST had no adverse impact on autobiographical memory(24).

The neurocognitive outcome differences between ECT and MST suggest that the former, but not the latter, may adversely impact the functioning of frontal and temporal brain structures (e.g., dorsolateral prefrontal cortex, hippocampal formation) and neurocircuitry(11, 59). Computational modeling and preclinical evidence demonstrate that the ECT-induced E-field is broad across the whole cerebral cortex with deep penetrance into both cortical and subcortical brain structures(60–63). Also, ECT, relative to MST, tends to stimulate a greater amount of brain volume(63) and its induced electrical field strength tends to be greater in both the frontal and temporal cortices(64, 65). Clinical evidence has demonstrated that the ECT-induced E-field, even with RUL electrode placement, is broad with penetrance across the contralateral hemisphere and into subcortical structures(46, 49). Clinical evidence has found a direct link between the ECT-induced E-field and outcomes(49) including neurocognitive adverse effects (e.g., verbal dysfluency) that is independent from the ECT-induced seizure(51, 66).

In stark contrast, computational modeling and preclinical model evidence has found that the MST-induced E-field is confined to superficial cerebral cortex with little to no penetrance into cortical and subcortical brain structures(60) and minimal induced electrical field strength in the temporal cortex(64, 65). As there is limited clinical evidence regarding the effects of MST on neurocircuitry(67), future research is warranted to examine the respective contributions and combination of the MST-induced E-field and seizure on both neurocognitive and clinical outcomes.

While there were differential outcomes between the ECT and MST treatment groups on objective cognitive measures, the outcomes on the subjective cognitive measure remained relatively the same from baseline to end in both groups. The finding of no change in the subjective cognitive function is incongruent with the worsened performance on cognitive measures in the ECT treatment group and the relatively modest improvement in cognitive performance in the MST treatment group. This incongruence is consistent with prior research that has found inconsistencies between ratings on subjective cognitive functioning questionnaires and objective cognitive performance in adults with MDD(68, 69) and those treated with ECT(70). Future research needs to ensure the measurement of both objective and subjective cognitive performance to further inform how they are similarly or differentially impacted by antidepressant treatment strategies.

The neurocognitive safety of MST has clinically meaningful and beneficial implications for clinical practice. The quick time to reorientation could lead to quicker recovery and discharge from the procedure suite. Moreover, patients may have a quicker return to independence with instrumental activities of daily living (e.g., managing finances, driving) and functional responsibilities (e.g., employment, home management)(54, 71). In addition to patients with severe depression, MST could have applications in vulnerable populations such as older adults with MDD and cognitive impairment, neurodegenerative disorders with intractable behavioral disturbances, catatonia, and children and adolescents with treatment-resistant MDD(72, 73). Additional research is needed to establish MST as a safe and efficacious treatment in such vulnerable populations.

The neurocognitive safety of MST may allow for the development of accelerated paradigms, similar to accelerated TMS protocols.(4) One study reported on daily application of RUL-UB ECT, which provides a framework upon which to build(74). Accelerated protocols may be useful for life-threatening illnesses in which rapid response is required (e.g., catatonia, suicidality)(75, 76). Another paradigm could involve the integration of MST with evidenced-based psychotherapeutic and behavioral interventions(77). The synergy from such an integration has the potential to maximize antidepressant benefits and assist with prolonging remission during the continuation/maintenance recovery phases(78, 79). One small scale study demonstrated the feasibility and usefulness of integrating MST with dialectical behavioral therapy to treat patients with borderline personality disorder and suicidality(80).

Strengths and Limitations

This study had several strengths. To date, this was the first comparison of the neurocognitive outcomes of MST with the form of ECT in modern practice that has the fewest cognitive side effects. Thus, the study design used an ECT paradigm that provided the most rigorous test of whether MST may exert cognitive advantages relative to ECT. Using the most cognitively safe version of ECT also increased the likelihood that the study would remain well masked, since both groups shared an expectancy of less cognitive difficulty. Also, the study used a neurocognitive battery that was sensitive to the effects of ECT, assessed multiple cognitive abilities with verbal, visual, and psychomotor tests, had available alternate forms to minimize practice effects, and it provided comprehensive information on brain and behavior relationships.

The study also had several limitations. As there was no placebo control group, it is possible that MST could have produced transient cognitive adverse effects relative to anesthesia-alone; however, these study findings are in line with other studies of MST(20, 81) that also found neurocognitive stability and possible neurocognitive improvement after completion of the treatment course. Furthermore, prior nonhuman primate studies have found neurocognitive outcomes post MST to not significantly differ from anesthesia-alone(18, 56). As this study only focused on acute neurocognitive outcomes (e.g., 24 – 72 hours post last treatment), there was no information provided regarding longer-term neurocognitive outcomes. Research has demonstrated the neurocognitive effects of ECT tend to be transient and resolve over time(82), and research is warranted to examine the longer-term neurocognitive outcomes of MST. It is possible that neurocognitive test performance could have been moderated by practice effects due to repeat testing in a close time frame and certain demographic characteristics (e.g., highly educated sample). These limitations were mitigated by the use of alternate forms where possible to minimize potential practice effects and demographic characteristics were included as covariates in the analytic models.

The study had limited data collection on preservation of the blinding and use of lorazepam as needed for the study. Regarding study blinding, multiple strategies (e.g., prepping patients for both treatments, playing the sound of both devices when treatment was administered, keeping the clinical and neuropsychometric raters away from the treatment room until treatment had concluded) were employed to maintain the study blind, though with the lack of systematic data collection on blinding it is unclear if the blinding was effective. The findings of no change from baseline to after treatment in both the ECT and MST conditions on the subjective cognitive function measure suggested that the blind may have been maintained for the study participants. Nonetheless, future research needs to include systematic measurement to document if blinding strategies were effective. The use of lorazepam as needed in the study was not systematically documented and it is possible that it could have impacted neurocognitive performance based on research in healthy adults(83). However, some research in adults with MDD has suggested that the use of benzodiazepine and antidepressant medications could help to improve neurocognitive performance(84, 85). Future studies need to systematically document benzodiazepine usage to include it in statistical models.

The MST and ECT conditions were unmatched in terms of dosage relative to seizure threshold; however, the dosage of MST used in this study matched ECT in terms of efficacy(24), thus the contrast of their neurocognitive outcomes is clinically significant and meaningful. While we computed E-field simulations of the ECT and MST conditions (e.g., Figure S1), the study did not include neuroimaging to estimate individualized realistic E-field exposure via realistic head modeling, which might be useful to investigate sources of individual variability in outcome. However, even lacking individual E-field stimulations, the stark differences in neurocognitive outcomes were significant enough to be seen at the group level. Future studies to examine individual differences within group are warranted, and are underway(86), that include such individualized realistic modeling to elucidate neuromodulation-induced changes in neurocircuitry and their corresponding neurocognitive outcomes(50).

Conclusion

Relative to UB-RUL ECT, our results support the superior cognitive safety of MST. This finding, along with our prior report that MST demonstrates similar antidepressant benefits as UB-RUL ECT, substantiate that MST is a safe and beneficial non-invasive brain stimulation antidepressant therapeutic. Future research is warranted to replicate these study findings and to continue the development of MST in larger samples to optimize its application in individuals with major depression(72) and those with other severe neuropsychiatric conditions (e.g., Bipolar disorder, psychosis)(87, 88). Our findings and those that will stem from ongoing international MST randomized clinical studies(86, 89) will provide useful information to translate this safe and beneficial treatment from research into clinical practice.

Biographies

Ms. Elisabeth Bernhardt. Ms. Bernhardt reports research support from the National Institutes of Health.

Dr. Zhi-De Deng. Dr. Deng is inventor on patents and patent applications on electrical and magnetic brain stimulation therapy systems held by the National Institutes of Health (NIH), Columbia University, and University of New Mexico. He is currently an employee of the National Institute of Mental Health. The opinions expressed in this article are his own and do not reflect the views of the National Institutes of Health, the Department of Health and Human Services, or the United States government.

Dr. Mustafa M. Husain. Dr. Husain reports research support from the National Institute of Mental Health, National Institute of Neurological Disorders and Stroke, Stanley Medical Research Foundation, Neuronetics, Inc., MagStim (equipment only), Brainsway, Inc., and NeoSync. Consultant to the Neurological Devices Panel of the Medical Devices Advisory Committee, Center for Devices and Radiological Health, Food and Drug Administration (FDA). Editorial Board Member of the Journal of ECT and the American Journal of Geriatric Psychiatry.

Dr. Sarah H. Lisanby. Dr. Lisanby is inventor on patents and patent applications on electrical and magnetic brain stimulation therapy systems held by the NIH and Columbia University, with no remuneration. The opinions expressed in this article are the author’s own and do not reflect the views of the National Institutes of Health, the Department of Health and Human Services, or the United States government.

Dr. Bruce Luber. Dr. Luber is currently an employee of the National Institute of Mental Health. The opinions expressed in this article are the author’s own and do not reflect the views of the National Institutes of Health, the Department of Health and Human Services, or the United States government.

Dr. Shawn M. McClintock. Dr. McClintock reports research support from the National Institutes of Health. He is a consultant to Pearson Assessment. He receives teaching honorarium from Duke University School of Medicine. He receives royalties from Guilford Press, Inc.

Dr. Vishal J. Thakkar. Dr. Thakkar reports no biomedical financial interests or potential conflicts of interest.

Dr. Richard D. Weiner. Dr. Weiner reports no biomedical financial interests or potential conflicts of interest.