Enhancement of Treatment of Delusions in Schizophrenia Through Neuromodulation

Purpose: The objective of this study is to enhance a form of nonpharmacological treatment of delusions in schizophrenia using neuromodulation, and confirm the effect using electroencephalogram (EEG) and behavioural measures (Whitman et al., in press; Whitman, Minz \& Woodward, 2013). Specifically, transcranial alternating current stimulation (tACS) will be used to improve the outcome of participation in metacognitive training (MCT) for the treatment of psychosis. MCT is a group-based program developed directly from current cognitive neuropsychiatry research findings on schizophrenia and psychosis. MCT shares knowledge gained in research labs to help individuals experiencing psychosis become more aware of the thinking patterns involved in their illness. The main purpose of the metacognitive training is to help people change the thinking patterns that cause delusions, thereby avoiding relapse into illness or reducing the impact of delusions (Moritz et al., 2014). tACS is a non-invasive brain stimulation technique in which a weak electrical current is applied to the head in order to change the likelihood of neural firing in a region of cortex under the stimulation electrodes (Antal \& Paulus, 2013). Hypothesis: Neuromodulation will enhance the benefit of the non-pharmacological treatment program, MCT, in the reduction of severity of delusions in schizophrenia. Justification: It is now clear that helping individuals with schizophrenia become aware of the thinking patterns that underlie delusions provides effective tools for countering these thinking patterns (Moritz et al., 2013; Mortiz et al., 2014). The MCT program for treatment of delusions involves experiencing situations where everyday interpretations of reality must be questioned. Empirical studies and meta-analyses indicate that MCT is effective, even up to three years after completion of treatment (relative to active control conditions and treatment as usual (TAU)). Despite these successes, no treatment of schizophrenia can be considered highly effective, and there is always room for improvement in effect size. Neuromodulation can provide this improvement by biasing pathological brain functioning towards a healthy state prior to MCT, enhancing the openness to questioning everyday interpretations of reality, and enhancing the effect of MCT treatment of delusions. tACS is almost entirely non-invasive and purely modulatory, changing the likelihood of neural firing in a region of cortex under the stimulation electrodes. It is now possible to modulate at different frequencies of oscillation, in accordance with how the healthy brain functions when interpreting reality. The proposed research involves application of oscillation-targeted neuromodulation protocols to bias brain functioning towards a state that promotes questioning everyday interpretations of reality, thereby reducing severity of delusions, and enhancing response to a training program for treatment of delusions in schizophrenia by raising insight about the thinking patterns that lead to delusional thought. The investigators plan to document the brain changes induced by neuromodulation using performance and EEG measures taken on a decision making task that involves interpretation of reality, and for which performance correlates with the severity of delusions. This task involves judging the likelihood that evidence matches a hypothesis, and it has been shown that this judged likelihood increases when patients are strongly delusional. The investigators have shown that this task state (i.e., evidence-hypothesis match) elicits an increase in alpha power in the task positive network, and a decrease in beta power the default-mode network (Whitman et al., in press). This suggests that maximizing the power of the alpha band in the task-positive network should bias the brain towards a healthier state for weighing evidence to test hypotheses about reality, and this can be confirmed with EEG and behavioral measures taken on the aforementioned decision making task, a heightened response to MCT, and a corresponding decrease in the severity of delusions. Research Design: Neuromodulation involving tACS will target increasing the power of the alpha band at dorsomedial prefrontal regions, and this is expected to also decrease the power of the beta band in ventro-medial regions. This will be confirmed by EEG recordings and source estimation (Whitman et al., in press; Whitman, Minz \& Woodward, 2013). MCT is a four week program with eight one-hour sessions. TAU is a four week waitlist control group. Sham neuromodulation will involve application of random patterns of low-grade currents to the same brain region as the neuromodulation conditions (Antal \& Paulus, 2013). Severity of delusions will be measured by the Psychotic Symptom Rating Scales in psychosis (PSYRATS) (Haddock et al., 1999). The behavioral measure of interest is the rating of the match between the evidence and hypothesis, which should decrease as delusions decrease. Statistical Analysis: Between-groups (neuromodulation/MCT vs. sham/MCT vs neuromodulation/TAU) analyses of variance (ANOVA) will be carried out on change scores computed on behavioral measures and severity of delusions, pre- and post MCT or TAU, under the hypothesis that decreases in the severity of delusion will be greater in the neuromodulation/MCT condition relative to the sham/MCT and neuromodulation/TAU conditions. Control-Phase Studies: DREAM Study, Hearing Brain Study and The Effect of Neuromodulation on Saccades Study. Each described below. Purpose for Control Phase Studies: The objective of these studies is to establish the effectiveness of the EGI GTEN system in modulating brain oscillations in the cortex of healthy participants by means of transcranial alternating current stimulation (tACS). In this control/pilot phase of our study, we aim to establish that we can induce changes in the power of a specific frequency band in targeted cortical regions with neuromodulation using the GTEN system, and we will assess whether doing so temporarily alters performance on simple cognitive and perceptual processes in healthy controls. This will be the first step towards translating our stimulation protocol to the patient population for our primary study of interest (tACS as an adjunct to metacognitive training for delusions in psychosis). Justification for Control Phase Studies: Transcranial alternating current stimulation (tACS) is a non-invasive method used to influence ongoing brain oscillations in humans by applying sinusoidal currents to the scalp (Antal \& Paulus, 2013; Vosskuhl, Strüber, \& Herrmann, 2018). Human brain oscillations, as measured with electroencephalography (EEG), range from 1 to 100 Hz and are referred to as delta (1-3 Hz), theta (4-8 Hz), alpha (8-12 Hz), beta-1 (13-30 Hz), and gamma (30-40 Hz) oscillations. The oscillations, generated either spontaneously or in response to external stimuli, are the result of rhythmic or repetitive electrical activity of cells, and are known to be involved in the communication of brain regions (Vosskuhl et al., 2018). The synchronization of rhythmical electrical activity leads to increased power in the frequency band of the rhythmical activity. For example, the presentation of an auditory stimulus ("auditory entrainment") at a frequency of 40Hz increases the power of the 40Hz brain oscillations in the auditory cortex ;this phenomenon is known as Auditory Steady-State Response (ASSR; Galambos, Makeig, \& Talmachoff, 1981; O'Donnell et al., 2013). Previous studies have mainly relied on systems that use two large electrodes on the scalp to modulate the underlying brain activity in the oscillation frequency range of interest. The GTEN has 256 potential stimulation sites, as well as advanced source localization techniques, allowing for precise planning and application of current via tailored combinations of electrodes according to the 'reciprocity principle' (Dmochowski, Koessler, Norcia, Bikson, \& Parra, 2017; Fernandez-Corazza, Turovets, Luu, Anderson, \& Tucker, 2016). The 'reciprocity principal' is a process enabling optimal targeting/stimulation of the brain sources of recorded EEG signal. The GTEN system has in-built software for this very purpose, allowing researchers to program electrodes with the dominant power of frequencies of oscillation, using a small subset of electrodes (4-7) to optimally target the brain regions of interest. This reciprocity principal is ground-breaking in its capacity to demonstrate causal relationships between brain and behavior: by targeting the very brain sources producing brain activity of interest, and applying tACS while observing the behavioral effects on performance (e.g., reaction time, accuracy), one can experimentally link cognitive processes to specific patterns of brain oscillations. Establishing that we are able to use these methods in this way will be an important precursor to our study of tACS as an adjunct to metacognitive training for delusions. Our work using tACS to bias ongoing network activity in the brain rests on being able to accurately target those networks producing the activity of interest, and document whether doing so temporarily changes performance on simple cognitive tasks. Our "pilot/control" phase with healthy controls will familiarize us with the GTEN system and enable us to optimize the parameters under which GTEN tACS produces short-term changes in cortical excitability and cognitive processes. These insights will inform our larger study focussing on enhancing metacognitive treatment of delusions in schizophrenia using neuromodulation To validate that the GTEN tACS is able to bias ongoing oscillations in targeted cortical areas in a biologically-relevant manner, an initial study in healthy controls ("The Hearing Brain") will investigate the degree to which GTEN tACS applied to the auditory cortex mimics and/or influences the ASSR that endogenously occurs during perception of an auditory tone. To understand whether GTEN tACS temporarily alters performance on cognitive processes by biasing dominant patterns of oscillations, a second set of initial studies in healthy controls will be run called the "Dynamic Reciprocal Electroencephalography and Modulation (DREAM") study. The DREAM study will apply the reciprocity principle while participants undergo simple cognitive and perceptual tasks. This data collection in healthy controls will include the methods we have already received ethics permission to use with clinical populations (e.g., see 'Enhancement of Treatment of Delusions in Schizophrenia through Neuromodulation Protocol'). Piloting these methods will serve as a verification that tACS can alter cortical excitability in the cortical regions that are implicated in delusional thinking, which we have already been granted permission to neuromodulate in our primary study of patients with schizophrenia. The pilot study will also serve as a preliminary investigation of whether tACS has detectable influences on cognitive tasks that correlate with psychopathology (e.g., delusions). One task involves judging the likelihood that evidence matches a hypothesis, and it has been shown that this judged likelihood increases when patients are strongly delusional. We also aim to study the behavioural effects of tACS in the "The Effect of Neuromodulation on Saccades Study." Instead of using complex behavioural tasks, we aim to use a simpler and better-understood behaviour response - the oculomotor response. In this study, we aim to use tACS to stimulate the areas involved in oculomotor control to affect eye movements. Participants' gaze will be tracked binocularly using the Eyelink 1,000 system (SR Research, Kanata, ON). The oculomotor response is a well understood behavioural response that is not complex in nature and would allow us to study the effects of tACS on behaviour, this would expand our knowledge on the behavioural effects of tACS. The rationale behind using tACS is that we can target a specific ongoing brain oscillation. In this proof-of-concept study, we plan to use tACS to stimulate frontal eye fields (FEF), which we expect to interrupt the natural network balance, and slow saccades. Thus, the overall objectives of this study are (1) to identify the brain oscillations in FEF underlying saccade preparation and execution, and (2) change activity in these networks through neuromodulation. Hypothesis: In the "Hearing Brain" proof-of concept pilot study, we hypothesize that auditory entrainment at 40 Hz as well as GTEN stimulation at 40Hz, but not sham stimulation, will increase the gamma band power in the auditory cortex. Furthermore, we hypothesize stronger gamma band power after GTEN stimulation compared to auditory entrainment. Regarding the "DREAM studies," we predict that that GTEN applied based on the reciprocity principle will alter participants' accuracy and reaction time while undergoing simple cognitive and perceptual tasks. In particular, regarding the pilot of the methods of the primary study( Enhancement of Treatment of Delusions in Schizophrenia through Neuromodulation), we hypothesize that neuromodulation involving tACS will target increasing the power of the alpha band at dorsomedial prefrontal regions and bias our behavioral measure of interest (the rating of the match between the evidence and hypothesis (see Figure 1 in 'Enhancement of Treatment of Delusions in Schizophrenia through Neuromodulation Protocol'). Regarding the "The Effect of Neuromodulation on Saccades Study," we predict the following: 1. Increased power of alpha oscillation in the FEF will be detected by hdEEG when saccades are execute 2. GTEN-induced alpha oscillations in the FEF will increase saccade latency. Research Design: Experimental Design: proof-of concept study "The Hearing Brain" We will use a between-subjects design to investigate the influence of auditory entrainment, neuromodulation, sham stimulation (control condition), and control frequency stimulation (active control condition) on the power of the gamma band (30-40Hz) in the auditory cortex. All participants will be assigned to either the Neuromodulation, Sham, or Active Control (control frequency) condition, with or without also undergoing the Auditory entrainment condition. Neuromodulation: For each participant, we will establish a threshold at which the participant first detects the stimulations (e.g., due to phosphenes or skin sensations). In line with other threshold protocols, we will start with tACS stimulation at 40 Hz and 1000 µA for 1 second and increase the amplitude of the stimulation stepwise by 250 µA to a maximum of 3000 µA (Zaehle et al., 2010). Participants will be asked to indicate the presence of any sensations they experience. Each participant's maximum stimulation intensity will be kept 250 µA below their threshold for experiencing phosphenes or skin sensations (whichever threshold is lower). During stimulation, tACS will be applied over the auditory cortex with a frequency of 40 Hz. For each stimulation, the current will be slowly ramped up to the participant's tailored intensity and ramped down at the end of the segment. The stimulation electrodes in the current proof-of-concept study will be those identified as showing the largest increase in gamma power in an auditory entrainment pilot study. Auditory entrainment: Participants will be listening to a sequence of tones presented at 40Hz. Sham: This condition is used to control for expectation effects. Similar to the neuromodulation condition, the current will be ramped up and ramped down; however there will be no stimulation in between. Control frequency: People in this active control group would be modulated on the same electrodes and with the same current intensity, but at a different frequency band. An active control group is needed in order to make sure the effects are due to the applied frequency of interest and not just due to the current itself. Rest: After each neuromodulation, entrainment, and sham trial there will be rest period of 50 seconds to measure resting-state EEG. Each condition ( modulation, sham, auditory entrainment) will be 50 seconds long and presented 7 times in blocks. In between the conditions there will be a 50 seconds resting-state measurement to assess the durability of interventions on brain activity. Experimental Design: proof-of concept "Dynamic Reciprocal Electroencephalography and Modulation" (DREAM) study We will use a between-subjects design to investigate the influence of neuromodulation based on the reciprocity principle vs. sham stimulation (control condition) vs. control frequency stimulation (active control condition) on reaction time and accuracy in simple cognitive and perceptual tasks. These tasks include the Stroop Interference Task, listening to/ reading non-offensive non-emotional words, and remembering words or visual arrays. We also plan to pilot tACS in the evidence-integration task which has been shown to correlate with delusional severity in psychosis. To illustrate the reciprocity principle method for this task, participants will be assigned to receive either tACS or sham neuromodulation. Neuromodulation involving tACS will target the dominant pattern of brain oscillations, which we expect to be the alpha band at dorsomedial prefrontal regions, and doing so is expected to also decrease the power of the beta band in ventro-medial regions (Antal \& Paulus, 2013). This will be confirmed by EEG recordings and source estimation (Whitman et al., in press; Whitman, Ward, \& Woodward, 2013) during the task used to produce Figure 1. Sham neuromodulation will involve application of random patterns or very brief low-grade currents to the same brain region as the neuromodulation condition (Antal \& Paulus, 2013). The behavioral measure of interest in this task is the rating of the match between the evidence and hypothesis (see Figure 1), which has been shown to correlate with delusional severity in psychosis. Experimental Design: proof-of concept study "The Effect of Neuromodulation on Saccades" The experiments proposed are standard saccade experiments, similar to the ones conducted by Chen Zhang et al. In the prosaccade task, each trial begins with a fixation point at the center of the screen that lasts for 800-1200 ms picked randomly from an exponential distribution. Participants will be instructed to fixate on the fixation point. The fixation point disappears and, after a time gap of 200 ms, a peripheral target appears randomly at 10° to the left or right of the central fixation point. Participants will be instructed to make a rapid saccade in the direction of the target. They have 1,000 ms to complete the saccade to the correct location. In the anti-saccade task, the procedure remains the same but the participants are instructed to make a saccade in the opposite direction of the target. In the memory-guided saccade task, participants will be instructed to maintain fixation, after which two targets will appear sequentially for 100 ms each in one of four quadrants around the periphery of the screen. Participants will be required to maintain fixation for additional time and wait for the go signal (disappearance of the fixation point). Then the participants are required to make two saccades as accurately as possible to the remembered locations of the targets in the same sequence in which they occurred. The prosaccade, anti-saccade, and memory-guided saccade trials will be interleaved in a block. The different types of tasks will be differentiated by different colors of the fixation cross. tACS Protocol: All the participants will perform three blocks of the above experiment: pre-stimulation block (pre), stimulation block (stim/sham), post-stimulation block (post). In the pre-stimulation block, participants will perform all the above tasks, simultaneously their EEG data is recorded and their eye-movements are tracked. Based on the EEG data, the exact montage for frontal eye fields (FEF) stimulation will be estimated for each participant. Also, the exact alpha frequency for stimulation will be estimated. Based on these parameters, half the participants will be stimulated in the stimulation block. The other half of the participants will act as controls will undergo a sham block. In both cases, the participants also perform the saccade tasks. The post block is similar to the pre-stimulation block. Data Analysis: Proof-of concept study "The Hearing Brain": Power Spectrum density (PSD) for the range of 1 to 50Hz will be calculated for each of the resting-state measurements and averaged for each of the conditions. Repeated-measures (entrainment vs. neuromodulation vs. sham vs. control frequency) analyses of variance (ANOVA) will be carried out for the PSD values of gamma band and the control frequency bands (alpha, beta), under the hypothesis that neuromodulation and entrainment but not sham will lead to increased gamma band power but not to differences in the control frequencies. Proof-of-concept "DREAM" study: Between-groups (neuromodulation vs. sham vs. control frequency stimulation) analyses of variance (ANOVA) will be carried out on change scores computed on behavioral measures (accuracy, reaction time) in order to assess whether tACS applied using the reciprocity principle to bias ongoing brain oscillations alters cognitive performance. Proof-of concept study "The Effect of Neuromodulation on Saccades": EEG preprocessing: The hdEEG data recorded using the GTEN system will be analyzed using MATLAB 2019b using in-house scripts. EEGLAB 13.6.5b (Delorme \& Makeig, 2004) functions were used to generate the topographical plots. The spectral analysis will be performed using Chronux 2.12 toolbox (Bokil et al., 2010). The EEG data will be preprocessed using a Statistical Correction of Artifacts in Dense-array Studies (SCADS) (Junghöfer et al., 2000), which removes EEG artifacts based on statistical parameters. Firstly, the data will be filtered using an FIR bandpass filter with stop band frequencies at 2Hz and 55Hz. The filter will be designed using the MATLAB based signal processing toolbox. Then the EEG data will be epoched to contain the entire trial. Then, electrodes contaminated with artifacts will be rejected by the SCADS algorithm. The parameters used for electrode rejection are - maximum amplitude, maximum gradient and standard deviation of signal amplitude in each epoch, and each electrode. Further, the EEG data will be re-referenced to the average signal. Montage estimation using EEG: The exact electrode montage and stimulation frequency to be used in the stimulation block, will be estimated based on the EEG data collected in the pre-stimulation block. The EEG data is preprocessed and epoched w.r.t the saccade onset. EEG data from -1000ms to 500ms from the saccade onset is used to generate epochs. The source of the preparatory activity before the saccade is estimated using Constrained Principal Component Analysis (CPCA) (Takane \& Hunter, 2001). CPCA generates a component for the saccade preparation. The electrode loadings for the component will be used as a montage for tACS stimulation. Also, the frequency whose phase encodes the saccade latency (For more details refer to the methods in Drewes \& VanRullen, 2011) will be used as the stimulation frequency. Eye-tracking data analysis: All the data analyses will be performed on MATLAB 2019b. The eye-tracking data will be analyzed using in-house scripts. The eye-tracking data is parsed into different types of epochs. Missing data corresponding to eye-blinks will be rejected from the analysis. Also, the trials in which the participant does not maintain fixation will be rejected. The gaze velocity is used to discriminate saccades from fixations. Then data from each trial is analyzed to check if the participant performed the correct saccade. E.g., if in an anti-saccade trial, the participant saccades away from the target, then the trial is assigned as correct; otherwise, it is assigned as an incorrect trial. Then, for all the trials, behavioral metrics such as saccade accuracy, latency, velocity, amplitude, and trajectories will be extracted. The online effect of stimulation (Stim - Pre) and the long-term effect of stimulation (Post - Pre) is checked for each behavioral metric and each of the saccade tasks. To test hypotheses, analysis of variance (ANOVA) will be carried out on the dependent measures - saccade latency, accuracy, velocity, amplitude, etc, with Sham/Stimulation as the between-subjects variable. Clinical Impact: These preliminary studies will be conducted in healthy controls order to establish that we can induce changes in the power of a specific frequency band with neuromodulation using the GTEN system. Further, we aim to establish that we are able to detect short-term changes in EEG signal following tACS and alter cognitive processes that are relevant to the experience of psychiatric illness. Conducting this pilot research will be essential in translating our stimulation protocol to the patient population.