Symptoms of depression and anxiety are common in older adults and are associated with poor outcomes and the risk of dementia. The prefrontal cortex (PFC) is crucial for emotion regulation. Poor PFC function may underlie subclinical depression and anxiety symptoms in older people, which could progress to clinical conditions. Neurofeedback training based on electroencephalography (EEG) or functional near-infrared spectroscopy (fNIRS) teaches individuals to self-regulate different aspects of brain activity and induce neurocognitive improvements. This proposed project will examine whether prefrontal EEG and fNIRS neurofeedback training programmes can enhance the mood and cognition of older adults with subclinical depression and anxiety.

Background:

Subclinical symptoms of depression and anxiety are common in older adults, with some estimates indicating that these symptoms are present in 10-52% of community-dwelling older adults. Some studies have shown that older adults with subclinical depression and anxiety are more likely than those with low levels of relevant symptoms to be diagnosed with affective disorders and mild cognitive impairment or dementia later in life. Thus, interventions for older people with elevated subclinical symptoms of depression and anxiety are crucial for preventing affective disorders and dementia late in life. During negative emotional experiences, the prefrontal cortex (PFC) plays a pivotal role in downregulating activity. PFC dysfunction may cause different mood and anxiety symptoms.

Neurofeedback training is a non-pharmaceutical neurorehabilitation technique that can potentially improve prefrontal function and enhance mental health and cognitive functions. This technique uses sensory feedback to teach individuals to self-regulate specific brain activities, with the goal of inducing long-term neuroplasticity and functional improvements. Traditionally, neurofeedback training has been conducted using EEG, and much research has applied such training interventions for the treatment of a variety of psychiatric disorders. In recent years, interest in using fNIRS to deliver neurofeedback training has grown. The underlying mechanism of such training with fNIRS is different from that of training with EEG. Compared with EEG, fNIRS has a lower temporal resolution but a higher spatial resolution and is more resilient to movement artifacts. In addition, one recent study showed that patients with social anxiety disorder had reduced anxiety symptoms after fNIRS neurofeedback training.

Research Plan and Methodology:

Design: This proposed project has been designed in accordance with current consensus on the reporting and experimental design of clinical and cognitive-behavioural neurofeedback studies. The participants will be randomly and equally assigned to one of three neurofeedback training groups: (1) sham, (2) EEG, and (3) fNIRS. Each participant will complete a neurophysiological assessment (1) before, (2) immediately after, and (3) 1 month after intervention.

Participants: 90 older adults without dementia will be recruited via advertisements at PolyU and NGOs. The inclusion criteria is: (i) age of 60-79 years; (ii) right-handedness as assessed using the short form of the Edinburgh Handedness Inventory; (iii) a moderate or higher score on at least one of the depression and anxiety subscales (but not necessarily both) of the Depression Anxiety Stress Scale-21 (DASS-21); (iv) no history of neurological or psychiatric disorder; (iv) no history of traumatic brain injury requiring hospitalisation; (vi) not currently using psychotropic medication; (vii) ability to read Traditional Chinese; (viii) normal or corrected-to-normal vision; and (ix) a score of at least 19 on the Hong Kong Montreal Cognitive Assessment (HK-MoCA).

The inclusion criteria that the investigators planned to use were based on those employed by prefrontal neurofeedback studies in mood or anxiety disorders. Conventionally, participants are selected based on a certain threshold of depressive or anxiety symptoms, neither cognitive nor brain dysfunction constitutes an inclusion criterion. Nevertheless, since variation in cognitive and PFC functioning levels may affect the treatment response, subsequent analyses will consider baseline cognitive and PFC functioning levels.

Study Procedures: Potential participants will first be subjected to a screening evaluation to assess eligibility. Eligible individuals will be invited to PolyU for assessment and training. The training will comprise 10 60-min sessions conducted within 4 weeks. Each session will include 25-min effective training time, for a total training time of 250 min, in keeping with recent recommendations. In addition, the participants will undertake 3 experimental tasks under simultaneous EEG-fNIRS recording and complete several questionnaires at 3 time points, as described in the 'Neurophysiological Assessment' section. Multiple studies have demonstrated that EEG, fNIRS, and neurofeedback training can be applied to older adults over 70, and even to individuals with dementia. Therefore, the investigators expect that older adults who are screened for dementia by the HK-MoCA will be able to follow both the assessment and training protocols.

Neurofeedback Training: During training, participants will be asked to follow the instructions on a computer screen. They will complete five rounds of training task. Each round starts with a 30-s rest phase followed by 4.5 min of self-regulation phase. During the rest phase, a fixation cross will appear onscreen, and the participants will be instructed to sit still and relax. During the regulation period, the participants will be asked to make the square change from white to black (i.e., an intrinsic social reward) but will not be given specific strategies. The darkness of colour will represent the increase in either frontal alpha asymmetry or frontal oxyhaemoglobin (HbO) asymmetry. The values at the moment will be compared against the 20-s pre-regulation baseline. In the sham condition, participants will receive visual feedback based on pre-recordings and/or other participants' recordings. Participants will undergo a 3-min rest period before and after each training session to track changes in resting-state brain activity within and across sessions.

During each training session, a cap adjusted to the participant's head size will be used to mount the EEG and fNIRS sensors. The hardware setup will be the same for all groups to ensure that both the participant and the experimenter are blinded. For EEG to be recorded by the ANT eego rt8 amplifier (ANT Neuro, Hengelo, The Netherlands), electrodes will be placed at Fp1, F3, F4, Fz, Fpz, Cz, GND (ground), lower VEOG, and on the two earlobes (references). Data will be collected at 2,048 Hz. For fNIRS to be recorded by the wearable OctaMon+ system (Artinis Medical Systems, Gelderland, The Netherlands), two sources, each surrounded by four detectors positioned approximately 3 cm apart, will be placed on the scalp such that the two channels near the cerebral fissure on each side of the hemispheres are surrounded F3 and F4. Data will be sampled at 50 Hz. Depending on the training group, frontal asymmetry in terms of the difference in alpha power (8-13 Hz) between F3 and F4 and the mean change in HbO concentration between the left and right PFC will be chosen as the target objective. For both real training groups, real-time data streaming will be performed using the Lab Streaming Layer and OpenVibe according to published guidelines.

Neurophysiological Assessment: A 1.5-h neurophysiological assessment will be administered at each of 3 time points (pre, post, and 1-month follow-up) to evaluate the effects of neurofeedback training. The participants will complete the DASS-21 (Chinese version) to measure their depressive and anxiety symptoms over the last week; the Hospital Anxiety and Depression Scale (HADS; Chinese version) to measure their signs of anxiousness and depression during the previous week; the Pittsburgh Sleep Quality Index to measure their sleep quality over the last month; the Satisfaction with Life Scale (Chinese version) to quantify their general life satisfaction; and Lawton Instrumental Activities of Daily Living Scale (IADL; Chinese version) to assess independent living skills. The participants will also complete three computerised tasks to assess different components of frontal cognitive function under simultaneous EEG-fNIRS measurements, using the same setup as neurofeedback training. At the first visit, the participants will also complete the HK-MoCA to screen for dementia. Immediately after the intervention, they will be asked whether they know their treatment group assignment to check the strength of blinding.

Each assessment task (eyes open, Emotional Stroop, n-back) proposed for this research will comprise a difficult and an easy condition. The eyes open test is used to let the machine measure the activation baseline when the participants open their eyes. It requires participants to maintain their eyes open for 3 minutes. The Emotional Stroop task is used to assess inhibitory control. Participants are shown photos of different emotions with unrelated traditional Chinese emotion names. They are asked to name the photos by emotion. It requires participants to inhibit their emotions lead by the wordings and react to the content of the photo. Differences in accuracy and mean reaction time (RT) and changes in the prefrontal HbO concentration and theta power between the two conditions will be the dependent variables. The n-back task is used to assess working memory. During the task, participants are shown a sequence of digits and asked to judge via button press whether the digit they are seeing is zero (0-back; easy) or the same as the digit they saw two trials before (2-back; difficult). Differences in accuracy and mean RT and changes in the prefrontal HbO concentration and theta power between the two conditions will be the dependent variables.

Data Analysis: In this project, the primary outcome measures are mood and anxiety measures (i.e., DASS-21 and HADS scores), and the secondary outcome measures are task performance and PFC measures, as well as other mental health measures. The outcome measures will be analysed according to the CRED-nf checklist. Linear mixed models with group (sham, EEG, fNIRS), time (baseline, post, follow-up), and condition (easy, difficult) as the fixed factors; and the subject as a random factor will be used to analyse the behavioural, fNIRS, and EEG data. The investigators expect that participants in the two real neurofeedback training groups will demonstrate significant improvements in mental health, cognitive function, and frontal lobe function at the post and follow-up assessments relative to the sham group participants. In addition, the investigators will evaluate differences in pre-post changes in mental health and cognitive functions between the two real training groups. Moreover, the investigators will examine the correlation between baseline cognitive and PFC functioning levels and the pre-post changes in DASS-21 scores to elucidate individual differences in the treatment response for each neurofeedback group.

Inclusion Criteria:

• (i) age of 60-79 years;
• (ii) right-handedness as assessed using the short form of the Edinburgh Handedness Inventory (Veale, 2014);
• (iii) a moderate or higher score on at least one of the depression and anxiety subscales (but not necessarily both) of the Depression Anxiety Stress Scale-21 (DASS-21), which has been shown to yield reliable and valid scores;
• (iv) no history of neurological or psychiatric disorder;
• (v) no history of traumatic brain injury requiring hospitalisation;
• (vi) not currently using psychotropic medication;
• (vii) ability to read Traditional Chinese text;
• (viii) normal or corrected-to-normal vision; and
• (ix) a score of at least 19 on the Hong Kong Montreal Cognitive Assessment

Exclusion Criteria:

• does not fulfill any of the above criteria

Personal information obtained for this study will be kept totally secret, and each participant will be assigned a unique serial number. We will properly secure their identity and privacy, as well as the personal data of all participants. The Confidentiality and Privacy Ordinance of The Hong Kong Polytechnic University shall protect and keep all medical data. Personal information will be stored in an encrypted file with just a code number to identify it. All personal data will be removed five years after the study project is completed. Research findings may be published in international academic publications or at conferences, but all personal information will be kept confidential.

During each training session, a cap adjusted to the participant's head size will be used to mount the EEG and fNIRS sensors. The hardware setup will be the same for all groups to ensure that both the participant and the experimenter are blinded to the training group.