Effects of brain training on brain blood flow (The Cognition and Flow Study—CogFlowS): protocol for a feasibility randomised controlled trial of cognitive training in dementia

Introduction Cognitive training is an emerging non-pharmacological treatment to improve cognitive and physical function in mild cognitive impairment (MCI) and early Alzheimer’s disease (AD). Abnormal brain blood flow is a key process in the development of cognitive decline. However, no studies have explored the effects of cognitive training on brain blood flow in dementia. The primary aim of this study is to assess the feasibility for a large-scale, randomised controlled trial of cognitive training in healthy older adults (HC), MCI and early AD. Methods and analysis This study will recruit 60 participants, in three subgroups of 20 (MCI, HC, AD), from primary, secondary and community services. Participants will be randomised to a 12-week computerised cognitive training programme (five × 30 min sessions per week), or waiting-list control. Participants will undergo baseline and follow-up assessments of: mood, cognition, quality of life and activities of daily living. Cerebral blood flow will be measured at rest and during task activation (pretraining and post-training) by bilateral transcranial Doppler ultrasonography, alongside heart rate (3-lead ECG), end-tidal CO2 (capnography) and beat-to-beat blood pressure (Finometer). Participants will be offered to join a focus group or semistructured interview to explore barriers and facilitators to cognitive training in patients with dementia. Qualitative data will be analysed using framework analysis, and data will be integrated using mixed methods matrices. Ethics and dissemination Bradford Leeds Research Ethics committee awarded a favourable opinion (18/YH/0396). Results of the study will be published in peer-reviewed journals, and presented at national and international conferences on ageing and dementia. Trials registration number NCT03656107; Pre-results.

neurovascular assessments at the CHiASm offices. Do these assessments occur at different times? Why would all testing not be on one occasion at CHiASM offices? P12,L12-Will ACE-III will be an outcome measure and a stimulus for CBFv changes? or will CBFv measurements occur while baseline cognitive data is being collected? Please clarify. If ACE-III will serve as both, but cognitive data will be collected separately to CBFv data, I would suggest a different cognitive tests to avoid practice effects, which would likely occur differentially between the three groups and introduce bias. P12,L22-The qualitative component includes questions about the testing periods. Please change "cognitive training program" to read "follow-up testing" P12, L25-Will you offer the participants the opportunity to participate in an interview or a focus group? I would suggest that interviews would be the best option. The participants who are most likely to experience the most problems due to cognitive impairment, are also the least able to contribute in a focus group situation. For these participants interviews with carers and PwD would probably be the best approach. It is also probably best to collect data in the same manner for all participants ie interviews. P14, L20-Given that you will be running a two-way ANOVA to look at the effects of intervention and population, are the three t-tests necessary? The ANOVA will determine the presence of effects and post-hoc analyses will be able to identify the individual comparisons which are significant. P15, L39-Please include one sentence about verbal assent also being obtained for people who do not possess the capacity to provide their own consent, in addition to the written consent of the personal consultee.

VERSION 1 -AUTHOR RESPONSE
Reviewer 1 I'd like to commend the authors on a very well written protocol and well thought-out novel study. I have no major changes but some small changes to help clarify the study for the reader and some queries/comments for the authors to consider. Being a feasibility study many of the changes suggested are less important for this study, but it is often useful to run pilots and feasibility studies as close to the protocol of the upcoming, full study to assess feasibility most accurately.
Comment 1 P5, L13-Did you mean "and"? cognitive and non-cognitive Response 1 Thank you, we have amended this to 'and/or' to reflect that patients can present with or without noncognitive deficits.
Comment 2 P5, L47-The authors have listed a training duration of 8-12 weeks? Will the training program differ in length for participants? Or is a final decision on the duration not yet made? Please clarify and justify if the training program will be of variable duration for participants.

Response 2
The duration of the programme was under development with Lumosity© at the time of submission and is now fixed at a 12-week duration. This duration was selected based on their previous research of sufficient "dose" of training for benefit. This has been amended throughout the protocol.
Comment 3 P5, L50-52 You provide a good summary sentence of cognitive training. Can you please provide a one sentence on each of cognitive rehabilitation and cognitive stimulation, just for completeness and differentiation?

Response 3
Thank you, this information has been added for completeness.

Response 4
Thank you this has been corrected to within.
Comment 5 P6, 37-A bracket was opened but not closed. Using "such as" may be better than brackets?
Response 5 This has been changed from brackets to 'such as'. An amendment has been requested to have only a lower cut-off score. The original aim for including cut-offs was to prevent patients with very severe cognitive function being inappropriately enrolled into the study. However, from a practical perspective many patients do not neatly fit into these cut offs as a result of varying degrees of pre-morbid cognitive function. In addition, the distinction between MCI and Alzheimer's disease is based on functional status, which is a clinical, not cognitive, assessment. Therefore, in order to improve the accuracy of diagnosis for enrolled participants we will enrol participants based on their clinical diagnosis, rather than cut-off scores, and these have been amended in the protocol.

Response 9
This has been amended accordingly.
Comment 10 P9, L41-Including anti-dementia medications is less of a problem for the feasibility study (where outcomes are less important), but this may represent a confound in the larger study and authors may wish to consider this for the larger study.

Response 10
Thank you for the comment. We agree that this would be an important confounder and would be a potential concern in a larger, fully powered trial. We did not feel it was ethical to stop these medications for a feasibility study, and practically speaking the majority of patients with a diagnosis of Alzheimer's disease will be established on these medications.

Comment 11
P9, L51-Considering that the intervention is based on computerised brain training, having sufficient computer skills may be an important inclusion criteria to add to the access to computer/internet.

Response 11
We are only enrolling participants with adequate technology available as listed in the inclusion/exclusion criteria, and as such, they will most likely be familiar with their own technology. Participants are provided with a demonstration of the cognitive training prior to starting and a user/troubleshooting guide. The programme itself requires minimal computer skills to be able to access and use it, and has been designed in a simple manner to avoid issues around lack of computer skills. In addition, a useful part of the feasibility assessment would be identifying this as a potential issue and how common it is.
Comment 12 P9-Exclusion criteria: Perhaps excluding participants who have used computerised brain training in the last 6 months?

Response 12
As the primary aim of this study is feasibility, and the trial is not powered to determine changes in effect sizes, this is of a lesser concern in this instance. However, this is an important point and something we would certainly consider incorporating into a larger, fully-powered trial.
Comment 13 P10, L8-Mixed dementia (Alzheimer's disease with elements of vascular dementia) is quite common. Will participants with known mixed dementia be excluded?

Response 13
At this stage we are only enrolling participants with Alzheimer's disease, and so participants with mixed dementia will be excluded. In future studies it would be beneficial to include this patient group, in addition to vascular dementia. The criteria quoted (NIA/AA 2011) will cover this as an exclusion.
Comment 14 P10, L34-5 x 30 minute sessions/week is a quite a lot more training than most studies. Even with less training than this compliance/dosage is often a complication. With this level of training I would anticipate problems with compliance especially with the impaired populations due to time constraints/fatigue. This would potentially result in the lowest rates of compliance in more impaired participants which would be problematic for interpretation of results. The amount of testing will be investigated in the qualitative component, it might be best to ask participants how much testing would be acceptable if you proceed with 5 x 30 minutes.

Comment 15
Thank you. We agree that this is a significant dosage. This was taken on the recommendation from Lumosity© concerning their previous dose effect research data. This is in part, one of the reasons for conducting a feasibility trial to investigate if this dosage is feasible and at which levels of impairment (i.e. healthy, MCI, or established dementia). As part of the trial we will be measuring compliance (conducted through the CT software). In addition, this information will be beneficial going forward to identify the most appropriate or benefiting group for cognitive training. Thank you for the suggestion, we will certainly include the compliance and dose under feasibility and barriers to training in the qualitative component of the study.

Comment 16
P10, L49-Could you please elaborate on how exercises will be selected? You also said that the training program would be personalised to the participant. By this do you mean that different participants will be doing different tasks due to how they perform in different domains? If so, without a very large sample size this would results in very "dilute" effects of the intervention. These effects are generally small, when detected. I think focussing on certain domains for all participants may yield better results (more important for the full study).

Response 16
The exercises are selected from Lumosity's© commercially available software. The exercises have been classified by Lumosity© under specific cognitive domains that they are intended to target and train. Thus, we are targeting five specific cognitive domains with the task-activation protocol and have selected tasks from Lumosity's© software which have been categorised under these domains. All participants will train on the same exercises, but at different levels depending on their cognitive ability. This is a design feature of the software so that it adapts to each participant. The information in this section has been updated to clarify this further.
Comment 17 P11, L19-The same investigator should not provide telephone support to the intervention group, as well as do follow-up assessments. This introduces a definite bias. Is it possible that different team members carry out these separate roles?

Response 17
Unfortunately, for this study we are resource limited to one investigator completing the assessments. This is in part due to the skills needed to undertake the neurovascular coupling protocol. As the primary outcome is feasibility, we feel this is an important concern and one which has now been highlighted in the limitation section. As a feasibility study, the bias is of a lesser concern, but certainly if this study is taken forward separate team members would be needed to complete assessments and provide support.

Comment 18
Could you please elaborate on the support that will be provided? Will this be only for participants who are experiencing problems or will all participants receive phone calls?

Response 18
The telephone support is mainly to troubleshoot issues occurring with technology as this is a primary concern for the feasibility study. Participants are being offered a weekly call, however, it is optional and so some participants may decline this if they feel it is too intrusive. As the controls are not undertaking any intervention we are not providing this service to them. We understand that ideally this would be the case to match the controls more closely to the intervention, however resources do not permit this number of weekly phone calls to be undertaken. This has been added to the limitations section.
Comment 19 P11, L32-Shouldn't the control group attend 3 testing sessions? Eg Wk 1. Intervention/control: baseline testing; Wk 8/12. Intervention: follow up testing, control: pretraining testing; wk 16/24. Control: follow-up testing. It might be helpful for the reader to include a diagram showing, intervention periods, testing periods, qualitative component and weeks that all of these elements occur for intervention & control?

Response 19
Due to time constraints, a cross over trial was not possible so the control group will only be attending two sessions. Only participants who have completed the training programme will be invited for interview after the follow-up assessment is complete. A flow chart has been provided in Figure 2.
Comment 20 P11, L36 & P11, L50-L36 you say that there are two assessment sessions (baseline and follow-up) that can occur at CHiASM offices or the participants' home, but then L50 you say there will be neurovascular assessments at the CHiASm offices. Do these assessments occur at different times? Why would all testing not be on one occasion at CHiASM offices?

Response 20
Ideally, all assessments will be completed at the same time at the CHiASM office. However, for participants with dementia who may fatigue more quickly, we have included the option for them to split the assessments across 2 visits if necessary. This is a common practice for dementia research studies within our locality. However, we anticipate the majority of participants will be able to complete all the assessments in one sitting. Again, this will be part of the feasibility assessment to understand level of burden the assessments place on participants with cognitive impairment.

Comment 21
P12,L12-Will ACE-III will be an outcome measure and a stimulus for CBFv changes? or will CBFv measurements occur while baseline cognitive data is being collected? Please clarify. If ACE-III will serve as both, but cognitive data will be collected separately to CBFv data, I would suggest a different cognitive tests to avoid practice effects, which would likely occur differentially between the three groups and introduce bias.

Response 21
The cognitive test scores will be collected at the same time as the CBFv measurements, thus cognitive tasks will not be repeated and will be compiled with the remaining ACE-III tasks to generate a full cognitive assessment. Therefore, there are no concerns for practice effects.

Comment 22
P12,L22-The qualitative component includes questions about the testing periods. Please change "cognitive training program" to read "follow-up testing"

Response 22
This has been changed accordingly.

Comment 23
P12, L25-Will you offer the participants the opportunity to participate in an interview or a focus group? I would suggest that interviews would be the best option. The participants who are most likely to experience the most problems due to cognitive impairment, are also the least able to contribute in a focus group situation. For these participants interviews with carers and PwD would probably be the best approach. It is also probably best to collect data in the same manner for all participants ie interviews.

Response 23
We agree that interviews would be most appropriate for patients with cognitive impairment and their carers. The focus group option was a recommendation taken on board by the PPI group to provide choice to patients with different preferences. We plan to offer the option of both to participants and identify which is the patient preferred option to take forward in future studies. We anticipate this will most likely be interviews, as suggested by the reviewer, given that they can be conducted more flexibly in the participants' home.
P14, L20-Given that you will be running a two-way ANOVA to look at the effects of intervention and population, are the three t-tests necessary? The ANOVA will determine the presence of effects and post-hoc analyses will be able to identify the individual comparisons which are significant.

Response 24
Thank you, the wording in this section is confusing and the t-tests are for differences in baseline demographics, and the differences in outcome measures are by ANOVA, with post-hoc tests. This section has been updated to clarify this.

Comment 25
P15, L39-Please include one sentence about verbal assent also being obtained for people who do not possess the capacity to provide their own consent, in addition to the written consent of the personal consultee.

Response 25
A statement to this effect has been added under sample selection. Blinding -in this feasibility study, neither participants nor researchers will be blinded, only data analysis will be performed blindly. Are there any creative ways to limit placebo effects and minimize the difference in experiences of the experimental and control groups? Will control group receive weekly support phone calls and some placebo activity? Also, will you monitor control group to ensure that they do not increase cognitive activity or begin training independently concurrent with the start of the trial?

Response 1
We agree that lack of blinding is a significant limitation to the study, and is a common problem for cognitive training studies where an adequate active control condition is still under debate. Unfortunately for this study we are resource limited to one investigator completing the assessments. This is in part due to the skills needed to undertake the neurovascular coupling protocol. As the primary outcome is feasibility, we feel this is an important concern and one which has now been highlighted in the limitation section. As a feasibility study, the bias is of a lesser concern, but certainly if this study is taken forward separate team members would be needed to complete assessments and provide support. The telephone support is mainly to troubleshoot issues occurring with technology as this is a primary concern for the feasibility study. Participants are being offered a weekly call, however, it is optional and so some participants may decline this if they feel it is too intrusive. As the controls are not undertaking any intervention we are not providing this service to them. We understand that ideally this would be the case to match the controls more closely to the intervention, however resources do not permit this number of weekly phone calls to be undertaken. This has been added to the limitations section.
Contamination from the control group is a risk with studies of cognitive training, however, as the control group are receiving the intervention for free (waiting listed) they are incentivised to remain in the study as such and not to obtain the training through other means (where they would invariably have to pay to access). We have implemented an additional requirement to all articles to include 'Patient and Public Involvement' statement within the main text of your main document. Please refer below for more information regarding this new instruction: Authors must include a statement in the methods section of the manuscript under the sub-heading 'Patient and Public Involvement'.
This should provide a brief response to the following questions: How was the development of the research question and outcome measures informed by patients' priorities, experience, and preferences? How did you involve patients in the design of this study? Were patients involved in the recruitment to and conduct of the study? How will the results be disseminated to study participants? For randomised controlled trials, was the burden of the intervention assessed by patients themselves?
Patient advisers should also be thanked in the contributorship statement/acknowledgements. If patients and or public were not involved please state this.

Response 1
This section has now been added to the methods in line with the guidance provided above and we have thanked the patient contributors in the acknowledgements section.
Comment 2 2. Please re-upload your supplementary files in PDF format.

Response 2
These have now been uploaded in pdf format.

Comment 3
3. Please remove all your figures in your main document and upload each of them separately under file designation 'Image' (except tables and please ensure that Figures are of better quality or not pixelated when zoom in). NOTE: They can be in TIFF, JPG or PDF format and make sure that they have a resolution of at least 300 dpi. Figures in DOCUMENT, EXCEL and POWER POINT format are not acceptable.

Response 3
The figures have been converted to PDF format separate from the original manuscript. They have clarified several aspects of the paper and I think it reads a lot more clearly. I look forward to the results of the full study in the future. I think there is only one paragraph that still requires some minor clarifications: -In the data collection paragraph could you please include a few words to indicate that CBFv measures will be obtained while participants undergo the cognitive testing? -Could you please include some of the information you have included in response 20? Eg assessments will be carried out at CHiASM offices but if participants become fatigued there is the option to complete this testing later at home.
There were two remaining instances where "8" needs to be removed from "8-12 weeks": "Data collection and Intergration" paragraph and "Methods and Analysis" section of abstract. Please state any competing interests or state 'None declared': None declared Please leave your comments for the authors below I'd like to thank the authors for their responses to my questions. They have clarified several aspects of the paper and I think it reads a lot more clearly. I look forward to the results of the full study in the future.

VERSION 2 -AUTHOR RESPONSE
I think there is only one paragraph that still requires some minor clarifications: Comment 1 -In the data collection paragraph could you please include a few words to indicate that CBFv measures will be obtained while participants undergo the cognitive testing?

Response 1
Thank you for the comment. We have added the following statement to clarify the exact measures that will be obtained: Data will be collected on the following parameters: average CBFv, ETCO2, HR, and BP at rest over five minutes, and peak percentage change from baseline in all parameters at 5-10 seconds, and 10-20 seconds after task activation for each of the five cognitive tasks.