Improving Visual Field Deficits With Noninvasive Brain Stimulation

Transcranial current stimulation Noninvasive transcranial current stimulation (tCS) has been safely used in human for decades. Noninvasive current stimulation techniques use battery-powered current generator devices that have a built-in circuitry to limit the current above a certain level, typically 2 mA (milliampere). tCS has been widely used during the last decade demonstrating non-significant risk to participants (Antal et al., 2017; Brunoni et al., 2011; Iyer et al., 2005; Nitsche et al., 2008; Nitsche \& Paulus, 2011) This study uses random noise (i.e. tRNS) which results in less net charge being applied than in tDCS (transcranial direct current stimulation). There is limited reporting of side effects from tCS using alternating currents (tACS) or random noise (tRNS) in the literature. Studies that have used tACS, have also reported adverse effects similar in nature to effects described in the tDCS literature, for example, headache, sensations under the electrodes and visual sensations (Antal et al., 2017; Antal et al., 2008; Brignani et al., 2013). Adverse effects that have been described in the tCS literature are described here to offer a conservative assessment of possible adverse effects. The most common side effects associated with tCS according to the most recent data available are: (Antal et al., 2017; Nitsche \& Paulus, 2011; Feurra et al., 2013)

Sensations reported by subjects under the electrode for tDCS (These sensations can sometimes continue throughout and for a brief period following completion of the tCS but usually resolve shortly after the initiation of tCS):

Mild tingling (20-70%); light itching (30-40%); slight burning (10-22%); discomfort or mild pain (10-18%)

Effects reported that occur only during tCS:

Visual sensation during switching on and off the stimulation (11%)

Other effects that can occur both during and after tCS include:

Moderate fatigue (35%); skin redness (30%); headache (10-15%); difficulties in concentration (11%)

Additionally, the following rare side effects have been described:

Nausea (3%); nervousness (<5%); ringing in the ear (<1%); hypomania has been reported in a few participants receiving tDCS for bipolar disorder and depression but never in normal controls (Subjects with a history of a psychiatric disorder will be excluded from the study) Although it has never been reported in tCS, seizures are a theoretical risk. A consensus paper supports that a tCS (including tRNS used in the present protocol) related seizure has never been reported in the literature, including studies conducted in older subjects and post-stroke subjects (Antal et al., 2017).

tRNS Visits The tRNS study visits will be conducted at BIDMC. Participants will be allowed to miss up to 15% of the visits. Additional sessions will be added on to reach the expected number of visits if it is within a reasonable timeframe as determined by the investigator.

Review of tRNS side effects and adverse events will be completed daily before and after stimulation. Any changes in medication or medical history will be assessed on a daily basis.

Set up for tRNS includes placing a cap and/or band with electrodes on the participant's head and applying gel underneath electrodes - stimulation will be initiated once the visual training program is set up and ready to be launched (or immediately if the participant is not in visual stimulation group). Stimulation or sham will be administered. This will last for 20 - 30 minutes. If the participant is in the visual training group they will perform the computer or VR based task during this stimulation/sham.

Within Group 1a and Group 2, half of the cohort will be stimulated with tRNS, and the other half will be sham-stimulated. Within Group 1b, the entire cohort will be stimulated with tRNS. The V1-lesioned brain hemisphere in VFD subjects and the homologous area in the healthy hemisphere will be targeted. For tRNS, 20 - 30 min of 1.0 mA current will be delivered to electrodes bilaterally positioned over O1/O2 (Herpich et al., 2019). Current direction will oscillate randomly within a high-frequency range (101-640Hz). For sham the same stimulation parameters will be used as in the active condition, except the stimulator will be programmed to turn off after a 20s ramp-up to 1.0 mA. With this "fade in" procedure participants report similar scalp sensation for both real stimulation and sham stimulation. All devices used for stimulation have "blind modes", where the investigator and participant are blinded to the type of stimulation.

Monitoring and Safety Plan Adverse effects will be collected from the start of the experimental protocol to the end of study participation. All adverse events, regardless of attribution to tRNS or pre/post assessments, will be collected and recorded using a standard adverse event form. Participants will be asked, in an open-ended way, about the presence of any such events daily. Intensity of each adverse event will be graded as mild, moderate, or severe. If an event occurs that is not expected (e.g. is not described in the research protocol), that indicates a change from baseline in cognition or vision, and/or requires immediate attention, such as a seizure, the study MD (or covering investigator) will be informed in real time to assess the event, advise on immediate care of the participant and to determine the necessary reporting steps. Any events that are serious or unexpected in nature, severity or frequency as compared to the risks described in the study plan will be reviewed by the principal investigator to determine the relationship of the event to the study. Reportable events will be submitted to BIDMC per determined policies.

A licensed physician, credentialed at BIDMC, will be available by pager during all tRNS visits at BIDMC. Furthermore, the person applying tRNS is trained to continually assess participants during sessions to monitor for discomfort, to identify early symptoms of syncope (e.g. sweating, pallor) and recognition of seizures. In addition, all staff are trained to apply basic measures to keep the participants safe. For example, if a participant experiences pre-syncopal symptoms or a syncopal event, immediate care will be provided to relieve the symptoms (e.g. they will be placed in a reclining position). In addition, the research nurses in the Center are available to assist with a rapid assessment of the participant, implementation of recovery measures and monitoring as needed.

Recruitment Stroke patients will be recruited from the Stroke Unit at the Beth Israel Deaconess Medical Center. An initial triage will determine the original level of visual field deficit in the acute phase by inspecting the participants' charts retrospectively and looking at the NIH Stroke Scale (NIHSS) items for visual deficits. Patients who present with visual field deficits and comply with the inclusion and exclusion criteria will be contacted and invited to participate.

Individuals interested in the study are asked to contact the Center for Non-invasive Brain Stimulation at the BIDMC. A Research Assistant will explain the aim and design of the study. If the participant is interested in the study, a telephone interview will be conducted to rule out some exclusion criteria. If the participant qualifies for the study, he or she will be invited to BIDMC where the study will be explained again in detail and the participant asked to carefully read and eventually sign the written consent form prior to entry into the study. The participant is encouraged to ask questions.

Sample Size and Cohort Splits This study is designed to test: (1) the usefulness of different visual tests, including typical psychophysical tests, in the evaluation of visual deficits after visual cortical damage in adults; and (2) the effect of visual retraining coupled with noninvasive brain stimulation in recovering visual perception after visual cortical damage in adult participants. Based on preliminary results, 92 participants with visual field defects (subacute and chronic ischemic strokes; chronic hemorrhagic strokes) will be enrolled. These numbers are based on a sample size calculation from published results, whereby the incidence of participants to respond positively was 60% (Herpich et al., 2019). It is anticipated that the study group will show a 75% incidence, with an alpha of .05 and a power of 80%. The estimated sample size is 78. However, given the potential high dropout rate (15%), 92 participants will be enrolled. Within Group 1 (Training + Stimulation), there will be 36 chronic and 10 subacute subjects. Within Group 2 (Stimulation only) there will be 36 chronic and 10 subacute subjects. Within all subgroups subjects will have a 50% chance of real vs sham stimulation. Subacute is defined as less than 6 months post stroke prior to entry into the study. Chronic is defined as more than 6 months post stroke prior to entry into the study.

Additionally 8 subjects with partial cortical blindness (subacute and chronic ischemic strokes; chronic hemorrhagic strokes) will be enrolled to use a virtual reality setup (Group 1b) which should provide power to detect a significant effect of brain stimulation on the primary outcome measure (direction difference thresholds). Sample size calculations suggest that 8 subjects are adequate to attain the goals of the project. The analyses will be within-subject tests that contrast pre- with post-training and pre- to post-stimulation performance in the same person. Two types of internal controls will also be used in each person - testing will be performed at untrained visual field locations, and at corresponding locations in the intact field of vision.

Statistical analysis The Student's t-test statistics and multifactorial ANOVA designs will be used to demonstrate significance of the effects. Based on similar experiments in animals and normal humans, and given the scientific goals, the sample size is appropriate and sufficient. The primary endpoint is: improvement in the motion discrimination task after training within the deficient visual field. Secondary endpoints are: (a) improvement in The National Eye Institute 25-Item Visual Function Questionnaire (NEI-VFQ-25); (b) reduction of the blind area in the visual fields as measured by Humphrey perimetry. Analysis will be performed using MATLAB. Data will be stored in the R drive at the BIDMC. EEG Data Analysis: Off-line inspection and removal of all EEG epochs with artifacts (e.g., eye blinks and eye movements) will be performed prior to averaging. There will be 60-100 repetitions of each condition with <15% rejected trials in all subjects. Averages will be computed for each subject for each electrode and each stimulus condition. Averaged responses will be used to identify waveform components of interest (P1, N1, N2, P2 and late peaks). Peak amplitudes and latencies of the N200 component relative to motion onset will be analyzed separately for horizontal (left, right), and radial (in, out) stimuli. Peak N200 amplitudes and latencies from all sites will be entered in mixed measures ANOVA designs with group as a between-subjects factor, and electrode site (e.g. Fz, FCz, Cz, CPz, Pz, Oz) as a within subjects factor. Greenhouse-Geisser adjustment for the degrees of freedom will be used for the recording site factor due to the inherent violations of the repeated measures assumptions of sphericity. Where appropriate, post-hoc analyses will be conducted using Tukey's HSD tests and a family-wise Type I error rate of .05.

For Group 1b (subjects using the virtual reality setup), we will first list the baseline characteristics of the experimental group of subjects. Continuous variables will be summarized by typical parameters such as mean, standard deviation, and range, and the normality of distribution will be determined using the Kolmogorov-Smirnov goodness-of-fit test. Categorical data will be summarized by frequency and percentage and analyzed using Fisher's exact test, as appropriate. Dropouts will be replaced to ensure the planned sample size of 8 participants is met. In general, analyses will follow the intention-to-treat (ITT) principle, unless otherwise specified. When the outcome variable is the change from baseline, participants will be included only if they have a baseline, the complete data from two-weeks of training, and one follow-up visit. We will look at the effect of Training +[active] tRNS vs performance in the good field at baseline. Analysis will be performed using custom scripts in Matlab and R.

Data safety and Auditing To safeguard confidentiality and privacy of protected health information, each study subject will be assigned a unique code number. A separate log linking the participant's name with study number and identifiers will be kept in a password-protected data file, accessible only by the study investigators. Names will not be provided to external sources other than the staff on the Center for Brain Science MRI protocol once the subjects have signed consent and agreed to be in the study if required. No identifying information will be published in which a participant could be distinguished. Data from this study will be entered into and stored in a secured drive available to investigators on the study behind the BIDMC firewall. All information needed at another center will be provided via secured email and/or secure file transfer.