fNIRS Neurofeedback in Highly Impulsive Participants With ROI Regions DLPFC and IFG

Aims:

The investigators will assess the effects - and specificity - of two fNIRS-based NF protocols, one targeting the DLPFC and one targeting the IFG. Both groups will also be compared to a "sham" training group with regulation of the fNIRS signal outside of the frontal lobes. Pre/post testing will be conducted in all three groups, i.e., measures of brain activity, cognition and impulsive behavior will be obtained before the start of the NF training phase as well as following the last training session. Thereby, a Go/Nogo task and an n-back task will be applied to evaluate training-related changes in brain activity and neuropsychological performance. The study population will be highly impulsive (healthy) individuals, who were previously found to show a reduced DLPFC and IFG activity (similar to patients with ADHD), and the investigators expect their DLPFC and IFG function to increase after fNIRS-NF training. Our hypothesis is that both active NF groups (DLPFC and IFG) will improve more than the placebo (sham training) group on all relevant measures. Thereby, specific advantages are expected in the DLPFC group for working memory data (n-back task) and in the IFG group for measures of inhibitory control (Go/Nogo task). Further, the investigators will test the effect of fNIRS-NF on the stress response using the TSST as a standardized stress test. It is hypothesized that fNIRS-NF training is also effective in reducing stress by strengthening stress-related regulation (that has also been related to the frontal lobe).

It is hypothesized that both trainings will be successful in reducing impulsive behavior; however, in a pre/post testing, specific effects of fNIRS-based NF of the DLPFC on working memory function and of fNIRS-based NF of the IFG on inhibitory control is expected. Correlations between both functions and impulsive symptoms will give an indication which training ROI may be more promising for the treatment of (specific subgroups of) ADHD. Correlations between regulation of different training ROIs will indicate the specificity of feedback regulation of circumscribed cortical areas.

Functional near-infrared spectroscopy (fNIRS):

An fNIRS system (ETG-4000, Hitachi Medical Corporation, Japan) will be used to measure hemodynamic responses within the prefrontal cortex (PFC).

All experiments will be conducted at the Department of Psychiatry and Psychotherapy, University of Tübingen (Calwerstraße 14/Osianderstraße 24, 72076 Tübingen). All medical-technical devices are CE certified and approved for application in patients and will only be used according to their approved purpose by persons who have the knowledge and experience to operate, apply and maintain the devices.

Neurofeedback:

One NF session comprises 2 feedback blocks each lasting 12 min and one transfer block lasting 8 min. A feedback block consists of 12 regulation trials. One regulation trial consists of 20 s resting time and 30 s regulation time. The task is to increase or decrease the hemodynamic activity in the DLPFC or IFG (or, for the sham control group, 4 channels of the temporal cortex, which is putatively not involved in impulsivity/ADHD symptomatology; furthermore, in an unpublished pilot study, these 4 channels were found to be inactive during a Stroop task involving cognitive control functions), with 50% activation trials and 50% deactivation trials, which were presented in random order. As a feedback of hemodynamic brain activation, a ball is shown to the participant on a screen, moving from left to right and depicting concentration changes in oxygenated hemoglobin by moving below or above a line in the middle of the screen. An arrow in the middle of the screen indicates if activation (pointing upwards) or deactivation (pointing downwards) is expected on a given trial. In activation trials, the concentration of oxygenated hemoglobin should increase in comparison to the baseline, in deactivation trials it should decrease. At the end of a successful trial (the ball moved in the expected direction for at least 7 s of the last 15 s regulation time), a rewarding feedback ("Gut gemacht, weiter so!", which is German for "Well done, continue like this!") is shown on the screen. A transfer block consists of 8 regulation trials in which the moving ball is not shown, but the reward at the end of the trial still indicates whether the participant was successful. The transfer blocks are included in order to facilitate transfer of the learned regulation strategies into everyday life.

Study population:

Participants will be recruited from the local student population via advertisements/email and from existing databases. An online system will be used to have potential participants fill out the ASRS in order to classify them as highly- or low-impulsive. During this online screening procedure, they confirm with a button press that they read the study information and the informed consent and that they agree with the usage of their data. They "continue" by a further button press and will be informed about their rights. Appropriate potential participants (highly-impulsive healthy subjects with ASRS scores >15 but <23) are contacted again for a study appointment. The complete study information, informed consent and data safety form will be sent to the potential participant at that time. When they arrive for the actual laboratory appointment, they will again receive explanations on all procedures and possible risks verbally and in written form. They are included in the study, when they give their written informed consent.

Sample Size Estimation:

The proposed project is planned as a randomized, controlled trial and comprises a total of 60 participants divided into an IFG-, a DLPFC and a placebo training group receiving either IFG-based, DLPFC-based or a sham (temporal cortex) NF control training. Based on the effect sizes of NF interventions typically observed in patients with ADHD (around d=0.8 for attention-related symptom improvements, as compared to different control groups including waiting list, EMG biofeedback and cognitive training; see e.g. Arns, de Ridder, Strehl, Breteler \& Coenen, 2009), subgroup sample sizes of n=20 would be the result of a formal power calculation, with a pre-defined α of 0.05 and an intended power of at least 80%. Closely related to the currently planned training protocol, recent own data confirm effect sizes of at least 0.8 for pre-post comparisons within a control group design. In detail, following eight training sessions with a frontal-lobe focused fNIRS-based NF training, highly-impulsive subjects showed positive changes in both target cognitive measures (standard deviation of RTs in a stop signal task) and cortical activation (left DLPFC) that were significantly stronger than in an EMG-based control training group (Hudak et al., 2017) with effect sizes of dcorr = 0.981 and 1.61, respectively (formula according to Klauer (2001) for pre-post control design studies). Taken together, and relying on the more conservative estimation of NF effects on the symptom level itself (see above), a sample size of n=20 per group and treatment arm thus seems to be adequate for the different aims and questions of the proposed study.

Risks and side effects:

FNIRS: No medical risks are associated with a fNIRS recordings; when laser diodes are used, which is the case in the present study, no significant tissue heating occurs. In order to avoid any accidental problems, all possible precautions will be taken. Guidelines and protocols on how to handle the measurement applications (e.g., comfortable adjustment of the NIRS cap) will be followed strictly as suggested by experienced laboratories. Moreover, the experimenters will work under close supervision of Dr. Ann-Christine Ehlis.

FNIRS-based NF: Risks associated with fNIRS-based NF treatment are few, rare, and quickly remediable. For individuals who are highly sensitive or susceptible, fNIRS-based NF may precipitate a migraine headache (muscle tension induced), hot flash, or mild anxiety (performance induced). Potential problems can be minimized and averted by obtaining an adequate medical history prior to treatment. Side effects are typically mild, transient, and quickly remediable allowing participants to continue with treatment.

Data Privacy:

All experimenters are bound to professional discretion. Medical doctoral students / master students potentially involved in the project will get detailed explanations of all procedures and will sign a contract for applying these guidelines. The confidential treatment of the data according to the data privacy act is guaranteed. As data is encoded on acquisition, it is not possible to assign published results to individuals. Recordings enabling a re-identification are locked and only accessible to staff members (for further details, see below). The following rules will be applied for handling the data.

Data acquisition and storage:

Data will be stored by using codes. These pseudonyms will be created for each participant (german: "Pseudonymisierung"). The codes are composed out of the name of the used paradigm and the running number of the participant.

There will be only three forms with the participants' name on them. First, participants will write their name, date of birth and signature on the consent form. Second, the experimenter will note down the participants name, date of birth and study code on the protocol form. This protocol will be the only way to link a participant's study code with her/his identity. Third, participants will sign a form for having received participation fee. All three documents will be stored apart from the other study documents in a lockable cupboard. The keys for this cupboard and for the room will be accessible solely with a licence signed by Dr. A.-C. Ehlis, head of the research group "Psychophysiology and Optical Imaging" at the Department of Psychiatry and Psychotherapy in Tuebingen.

The psychometric questionnaires will be stored in a cupboard in a lockable room. Only participants' study code is noted down on these questionnaires. The keys for the room will also be accessible solely with a licence signed by Dr. A.-C. Ehlis. These documents will be destroyed after ten years. Furthermore, the questionnaire data and participants' study codes will be entered into a data matrix of the programs Excel or SPSS for further analyses. The imaging data are also labelled with participants study codes.

The online questionnaire will be constructed and distributed using SoSciSurvey software. SoSciSurvey saves all data on a German server (located in Munich). SoSciSurvey does not share any data with third parties, does not save IP addresses and deletes all data if the project manager has not logged into the system for three months. All data are encrypted during data transfer. Thus, SoSciSurvey offers a very high security standard. Following the online assessment period, the data will be downloaded and stored on a password-protected hard disk of the University Hospital Tübingen (Psychophysiology \& Optical Imaging group). All data will then be deleted from the SoSciSurvey server.

Data analysis:

The coded data will be analyzed and results will be published in specific scientific journals. There will be no possibility to deduce information about single participants from these results.

FNIRS data will be analyzed with the software package implemented in the ETG-4000 (Topo software by Hitachi Medical Co., Japan) as well as MATLAB (The MathWorks Inc.), a commercial software applied to the offline analysis of the change in haemoglobin concentration. Additionally, statistical analyses will be conducted with the software packages Excel (Microsoft inc.) and SPSS. For statistical analysis, mainly ANOVAs for repeated measures and correlation analyses will be applied. Post-hoc comparisons will be conducted using t-tests for matched or independent samples as well as correlation analyses.