The relationship between individual alpha peak frequency and clinical outcome with repetitive Transcranial Magnetic Stimulation (rTMS) treatment of Major Depressive Disorder (MDD)

Highlights

• There was no relationship between outcome and individual alpha frequency (IAF) in patients treated at multiple frequencies.

• Post-hoc analysis showed a significant correlation between IAF measures and clinical outcome for patients treated at 10 Hz.

• The highest IAF quartile of 10 Hz-treated subjects had significantly greater clinical improvement than the lowest quartile.

• These findings indicate a relationship between endogenous oscillations and clinical outcome that differs across patients.

• Research should examine how endogenous oscillations may guide selection of optimal stimulation frequency for each patient.

Abstract

Background

The individual α frequency (IAF) has been associated with the outcome of repetitive Transcranial Magnetic Stimulation (rTMS) treatment of Major Depressive Disorder (MDD), but the association has been inconsistent.

Hypothesis

Proximity of IAF to the stimulation frequency, rather than the value of IAF per se, is associated with outcome for patients receiving 10 Hz rTMS.

Methods

We examined the relationships between IAF, rTMS stimulation frequency, and treatment outcome in 147 patients. All patients initially received 10 Hz rTMS unilateral treatment delivered to left dorsolateral prefrontal cortex (DLPFC) (10UL), with subsets of patients changed to unilateral 5 Hz to left DLPFC (5UL) or sequential bilateral (SB) stimulation (10 Hz/1Hz) to left and right DLPFC based upon worsening symptoms with or intolerance of 10UL. Outcome was percent change in total score on the Inventory of Depressive Symptomatology – Self Report (IDS-SR) scale from pre-treatment baseline to the 30th treatment. IAF values and absolute difference between IAF and 10 Hz (|IAF-10Hz|) were examined in relation to outcome for the overall sample and for each stimulation group separately.

Results

There was no correlation between IAF value, or |IAF-10Hz| and outcome in the overall sample. ANCOVA showed a significant interaction between IAF measures and treatment type. Post-hoc analyses revealed that IAF and |IAF-10Hz| were both significantly associated with degree of improvement (IDS-SR % change) for patients who received 10UL (P < 0.01) but not 5UL or SB stimulation. There was a trend-level difference in IAF between responders and non-responders only within the 10 Hz group, but not within the other treatment groups (n.s.). For the 10UL group, membership in the highest IAF quartile was associated with significantly greater clinical improvement than membership in the lowest IAF quartile (p = 0.0034).

Conclusions

IAF measures were associated with clinical outcome of patients treated with 10UL but not 5UL or SB rTMS treatment. This suggests that interactions between endogenous frequencies and treatment outcome may be related to the selected stimulation parameters and/or physiologic and clinical characteristics of patients who benefit from those parameters.

Introduction

Repetitive Transcranial Magnetic Stimulation (rTMS) for treatment of Major Depressive Disorder (MDD) is commonly administered to left dorsolateral prefrontal cortex (DLPFC) at a frequency of 10 Hz [1]. One possible effect of rTMS is the entrainment of oscillations in underlying cortex to the frequency of stimulation [2]. This change in oscillations may rapidly spread through brain networks to related brain regions [3,4]. 10 Hz is the center of the alpha (α) frequency band, with oscillations in the α band representing a thalamocortical rhythm [5]. Entrainment of α oscillations has been hypothesized to “reset” thalamocortical oscillators and may be related to the therapeutic mechanism of rTMS [6,7]. In healthy control patients, α oscillations play a key role in organizing the activity, function, and flow of information within resting state networks (RSNs) [[8], [9], [10]]. Fluctuations in the local functional Magnetic Resonance Imaging (fMRI) BOLD signal are strongly related to spontaneous α oscillations that largely explain evoked fMRI response variance [[11], [12], [13], [14]]. A possible role for α rhythms in depression is also suggested by findings demonstrating that α band oscillatory activity is significantly disturbed in MDD [15] and that α-band metrics early in the course of antidepressant medication (but not placebo) are statistical predictors of subsequent response or remission [16].

Each brain network has a preferred resonant frequency [17], which is best defined by its peak frequency [18]. While the range and peak frequency of α oscillations vary across individuals [19], individual alpha frequency (IAF; the single largest oscillatory peak in the α band) is a highly stable neurophysiological trait marker with high reliability observed across multiple measurements for up to six months in healthy adults [20], and stability over a course of standard rTMS treatment confirmed in a clinical sample [21]. It has been hypothesized that the IAF for each person may represent a resonant frequency at which networks are best engaged by neuromodulation treatments [6,7]. This hypothesis is supported by the finding that low intensity transcranial alternating current stimulation (tACS) of a circuit at its IAF upregulates α oscillations and may enhance spike-timing dependent plasticity (STDP) [22]. Applying principles arising from the relationship between stimulus amplitude and resonance frequency as characterized by the “Arnold Tongue” model [23], lower intensities of noninvasive brain stimulation may be able to entrain alpha oscillations when stimulation frequency more closely approximates the subject's IAF. The potential efficacy of this approach was supported by a subsequent multi-site randomized sham-control trial comparing low-field IAF stimulation relative to sham, albeit only in the per-protocol population which were treated at the correct IAF value and completed at least 80% of scheduled treatment sessions [24].

Clinical investigations have yielded conflicting evidence regarding IAF-guided rTMS. One study reported that a single session of rTMS at IAF +1 Hz was associated with greater enhancement of cognitive task performance than with stimulation at slower or faster frequencies [25]. A controlled trial of rTMS to DLPFC regions for treatment of negative schizophrenia symptoms showed that stimulus frequency matched to IAF was superior to sham, 3 Hz, or 20 Hz stimulation [26], and a subsequent schizophrenia trial confirmed IAF stimulation was superior to sham, independent of the stimulation target [27]. However, IAF +1 Hz did not subsequently prove superior for guiding rTMS treatment frequency in depressed patients when compared to outcome previously obtained with standard 10 Hz stimulation [28]. A follow up study observed a relationship between IAF and rTMS treatment outcome [29], although this was not replicated in a larger cohort of 106 patients [30].

We have hypothesized that differing results among prior studies may indicate that the difference between the intrinsic oscillations (IAF) and the stimulation frequency, rather than IAF alone, may be a key factor that determines the outcome of treatment. We thus examined the relationship between the absolute distance between IAF and 10Hz, as well as IAF values, and treatment outcome in 147 patients undergoing clinical rTMS treatment for MDD at two clinics. All patients were initially treated with 10 Hz rTMS stimulation administered to left DLPFC; 68 continued with this protocol (10UL). Subsets of patients were switched to either 5 Hz stimulation to left DLPFC (5UL, N = 39) or to sequential bilateral stimulation (10 Hz at left DLPFC followed by 1 Hz at right DLPFC, SB, N = 40) for the majority of their treatment sessions because of worsening symptoms or inability to tolerate 10UL. We tested the hypothesis that the distance between patients’ IAF and 10 Hz would be significantly related to treatment outcome for patients who received 10 Hz stimulation, but not for patients treated with the other stimulation protocols.