Data – all information pertaining to, but not limited to, the following: demographics, clinical, family history, neuropsychological, neuroimaging and biofluid measures. This includes the raw data and data derived from analyses of clinical, neuropsychological, neuroimaging and biofluid samples and measures.

Biospecimens – samples of DNA, RNA, plasma, serum, CSF, PBMC, fibroblasts, etc., and any products derived from these samples (including but not limited to proteins, induced pluripotent stem cells, etc.).

ALLFTD will follow a policy covering access to data with the intent of publication that acknowledges different levels of involvement.

Level 1 analyses are those that are specified in the specific aims found in the original application to the NIH. The principal investigators will be responsible for specifying the analyses and writing the manuscripts that relate to these specific aims. These investigators may designate a colleague to take the lead, but the ALLFTD Executive Committee must approve such a transition. The timing of Level 1 manuscripts will be left to the discretion of the ALLFTD Executive Committee.

Level 2 analyses may be proposed by ALLFTD co-investigators so long as they are outside the specific aims of ALLFTD. ALLFTD co-investigators may nominate a colleague or trainee within their team as a leader of such analyses or perform the analyses themselves. Manuscripts from Level 2 analyses require approval by the ALLFTD Executive Committee prior to submission for publication.

Level 3 analyses are ones that are proposed by qualified researchers who are not directly involved in ALLFTD. Proposals that fall under level 3 will be reviewed by the ALLFTD Executive Committee to ensure they do not conflict with ALLFTD aims or concurrent ALLFTD publication.

I request access to data collected by ALLFTD (ARTFL-LEFFTDS Longitudinal Frontotemporal Lobar Degeneration (ALLFTD) study, henceforth referred to as “Study”) for the purpose of scientific investigation, teaching or the planning of clinical research studies.

Prohibited use of AI Tools: This ALLFTD Data Use Agreement explicitly prohibits any individual who has obtained Study data from sharing raw data, specifically participant-specific data, with any other individual or entity, in any forum, in any medium. For a variety of reasons, including the important need to protect the privacy of research participants, all data must be directly released to requesting individuals from our Data Management Core.

The use of AI tools, including generative and analytical models, pose a genuine risk of inadvertent data sharing due to the nature of how AI tools process, store, and regenerate information. AI tools, including but not limited to those with public-facing interfaces, such as OpenAI's ChatGPT, usually do not offer guarantees regarding the containment of data inputs. Any data input into a prompt or output generated from a prompt could result with proprietary data infringed, sent, saved, viewed, misinterpreted, or used in unforeseen public ways. As a result, the use of such AI tools which may release Study participant level data to others, is in direct violation of ALLFTD Data Use Agreement policies.

However, an important exception to the above stated prohibition would be AI tools, including generative and analytical models, which explicitly prevent sharing data with others. Such AI tools may be developed and internally trained by individual research groups or academic institutions, and strictly used for the individual research group or academic institution provided that "sharing" data beyond the tool is strictly prohibited. Therefore, Study data may be analyzed using AI tools which provide guarantees that all data is safeguarded, contained, and will not be released to others.

Note that this amendment does not impose any new restrictions on the use of Study data, but is intended for clarification regarding the agreement to not further disclose these data beyond the uses outlined in this agreement and data use application and understand that redistribution of data in any manner is prohibited.

As of the writing of this Appendix the term "Artificial Intelligence" or "AI" means a broad range of technologies and tools that can generate and regenerate content in various mediums including text, images, videos, and audio, based on the content or data input to generate output in any medium response. The recent commercialization of these AI tools represents a significant challenge in maintaining the privacy of participants in the Study, and of participants in human subjects' research in general. Uploading Study data to a third-party platform that is not explicitly in compliance with the prohibition and restrictions on redistribution of Study data set forth herein is a clear violation of this Agreement.

ALLFTD requires investigators to adhere to the following terms when deciding whether a project utilizing these tools is in compliance with the terms of this Agreement:

• Computational resources

  • The application of any statistical or other analytic methods to Study data - whether branded as 'AI' and "AI Tool" or otherwise - should not be performed using computational or AI resources owned by third parties who engage in the long-term retention of user content that is either shared with other parties or used in the training of public-facing models, such activity could include the contents of the Study data set. This includes the use of third-party platforms for inference (e.g. the inclusion of Study data as part of an input or 'prompt' for a generative language model), training (including both the training of foundations model and the fine-tuning of pre-trained models via transfer learning or some other process), prediction, or any other task that could result in Study data being retained, shared, regenerated or otherwise by entities that are not in compliance with the DUA.

  • Some commercial entities may offer an 'opt-out' clause in the terms of use for their models, allowing users to decline the right of the entity to retain the data for training and other purposes. This does not constitute sufficient protection, and under no circumstances should this be viewed as an acceptable safeguard for sharing Study data with these third parties.

  • Investigators making use of remote computational resources such as Microsoft Azure, Google Cloud, Amazon Web Services, Runpod, or other computational services for the purposes of training AI models must take care to review the terms of use for the service to ensure that their workflow is secure.

• Public release of models trained using Study data

  • The use of Study data in training models is both permitted and encouraged, provided that the process is carried out in compliance with the DUA as outlined above. However, investigators who intend on making the weights of a trained model publicly available should consider whether the model could be used to easily reconstruct parts of the Study data set. The Study prohibits release of participant level data, even if it is sufficiently processed or altered and lack Study subject codes. All participant level data generated by the Study can only be released by the Study. Individual investigators who have generated participant level data based on Study data can request for their data to be shared by contacting Dr. Boeve at bboeve@mayo.edu.

  • This precaution is particularly important in the case of large language models and other highly overparameterized generative models. Numerous studies have demonstrated that it is possible to extract training examples from these large and complex models, and investigators should keep this fact in mind.

I understand that failure to abide by these guidelines will result in the termination of my privileges to access ALLFTD data. Furthermore, ALLFTD reserves the right to pursue damages for actions which violate ALLFTD guidelines.

After a request is approved, de-identified data will be made available to investigators to conduct analyses. All analyses will be based on data sets that have been prepared, cleaned and frozen on a quarterly basis.

All data requests should be submitted online using the request portal. This standardized application process asks the prospective authors to specify the principal hypotheses, the variables needed, and the analytic plan.

All data requests will be reviewed using the following criteria:

• Scientific merit and feasibility

• Appropriateness of the investigator’s qualifications and resources to protect the data.

• Appropriateness to ALLFTD goals/themes (if an ALLFTD investigator)

• Potential overlap with ongoing analyses

Biospecimens (blood, blood products, and cerebrospinal fluid) from ALLFTD participants are a scarce commodity and will be released to co‐investigators in a manner that parallels the levels of hierarchy described above. Requests for biospecimens will be reviewed by the ALLFTD Biospecimens Committee. Biospecimen samples will be distributed through the National Centralized Repository for Alzheimer’s Disease and Related Dementias (NCRAD). An MTA will be required for all biospecimen distributions; the MTA will specify requirements for returning results, proper acknowledgment, and any biospecimen‐specific procedures. Biospecimen requests may be rejected despite scientific merit if the distribution would substantively deplete the available samples.

Request ALLFTD Biospecimens here.

MR images were obtained for a subset of ARTFL and LEFFTDS participants, and will be obtained on all ALLFTD longitudinal participants. ALLFTD intends to provide MR images to interested investigators; however, as of mid 2020, this image sharing is on hold while additional deidentification procedures are implemented. Please visit ALLFTD Resource Sharing for more information.

If a data request is approved, the requestor must agree to:

• ALLFTD Investigators (Level 1 and Level 2):

  • Prepare a manuscript in a timely manner, as determined by the lead author and the ALLFTD Executive Committee

  • Upon request, share preliminary analysis results either via email or on a publications call

  • Agree to follow ALLFTD Authorship policies

  • Submit the manuscript to the ALLFTD Executive Committee and Data Management and Statistics Core Investigators 4 weeks prior to submission

  • Upon approval from the ALLFTD Executive Committee and DMS Core Investigators, submit the manuscript to all listed co-authors for review 2 weeks prior to submission

  • ALLFTD Executive Committee must be notified upon acceptance for publication

• External Investigators (Level 3)

  • Agree to follow ALLFTD data use policies and data display guidelines to protect participant confidentiality.

  • Acknowledge ALLFTD as data source.

  • May be requested to provide results of analyses to ALLFTD Executive Committee.

  • ALLFTD Executive Committee must be notified upon acceptance for publication

The ALLFTD Executive Committee reserves the right to require changes in the manuscript to avoid substantial conflict or overlap with other publications and to ensure proper description of informed consent, approach to confidentiality, acknowledgements of ALLFTD investigators and funding sources, and disclosure of potential and actual conflicts of interest.

All precautions to ensure confidentiality must be taken by recipients of ALLFTD data. The final data set will be stripped of identifiers prior to release for sharing and be transferred only with encryption and password protection by the ALLFTD Data Management Team. The code linking a subject’s identity to data will be maintained in a secure place and will only be accessible to research staff on a need to know basis. When known, exact genetic mutations will not be recorded in the National Institutes on Aging (NIA), National Institute of Neurological Diseases and Stroke (NINDS), National Cell Repository for Alzheimer’s Disease (NCRAD), database of Genotypes and Phenotypes (dbGaP), Central Neuroimaging Data Archive (CNDA) databases nor will it be entered into ALLFTD online electronic data capture system.

A parallel database to the ALLFTD electronic data capture system will be used to track and record genotyping data. Separation of this sensitive data is necessary to prevent accidental disclosure of participant mutation status to a member of the research team. Distributed research data sets will be coded with a second unique identifier (which is different from the study ID) to limit the risk of loss of confidentiality. A Global Unique Identifier (GUID), which is a randomly generated de-identified code unique to each participant will be generated for each participant and may be used to link de-identified data sets. There is always the possibility of deductive disclosure of participant identity because participants are limited to specific institutions, and the data set contains some demographic information, as well as detailed prospective information about their disease and mutation status, living situation, etc. Thus, we will make the data and associated documentation available to users only under the following prerequisites:

• Recipient of data will provide assurance of ability to secure the data set in accordance with the most stringent protections possible compliant with local IRB and Health Insurance Portability and Accountability Act (HIPAA for US sites) standards for such sensitive data.

• Recipient of data will provide a signed code access agreement for data usage – code access agreements are a simple statement that the recipient of the data will use the data only for research purposes and will not attempt to identify any individual participant.

• Recipient of data will guarantee that mutation data will be destroyed when analyses are complete.

Collaborative engagement is a key to deciding upon authorship. The following are the ALLFTD Authorship Policy Guidelines:

• The person who generate the first draft of the manuscript and takes principle responsibility for crafting the final version should be the first author on any ALLFTD publication.

• The senior author(s) should be the principle investigator(s) for all manuscripts from Level 1 and 2 data analyses.

• All named co-authors must meet appropriate standards for authorship, see below.

• Invite primary ALLFTD Investigators to participate as named co-authors. A list of these investigators will be provided upon request to any corresponding author.

• All publications from ALLFTD investigators or collaborators based on ALLFTD data must also include “on behalf of the ALLFTD Consortium,” as an author, which allows for the association and pubmed indexing of authors within the ALLFTD consortium. This is the most inclusive authorship list including leadership, site PIs, core leaders, and other individuals identified by the site PIs.

• All publications based only on ARTFL and/or LEFFTDS data please contact the ALLFTD Executive Team for guidance.

Standards for co-authorship:

• Must make substantial contributions to the study data.

• Must make meaningful contributions to the intellectual content of the manuscript during analysis, drafting, or revision.

• Must acknowledge participation as a named co-author.

• Must complete journal authorship forms in a timely fashion. Repeated delays in completing these forms will jeopardize co-authorship.

Abstracts: In many meetings, abstracts are often featured in press releases and thus might get wide media and professional attention. ALLFTD Abstracts must be cleared by the ALLFTD Executive Committee, just like full‐length manuscripts. Authors must submit abstracts to the ALLFTD Executive Committee at least 1 week in advance of the abstract due date. The ALLFTD Executive Committee reserves the right to require modifications to the abstract, which might include: description of informed consent, proper approach to confidentiality, proper acknowledgements of ALLFTD investigators and funding sources, reducing substantial overlap or conflict with other ALLFTD manuscripts, and ensuring proper disclosure of potential and actual conflicts of interest.

Other Personnel as Authors: ALLFTD investigators may include trainees or other site personnel as co‐authors provided they meet standards for authorship as defined above.

ALLFTD data sets include participant data collected during ARTFL & LEFFTDS. All policies on this web page apply to data collected as part of ARTFL & LEFFTDS. ARTFL & LEFFTDS investigators and grants must be acknowledged in any publication utilizing data from either/any study.

All data sets distributed prior to April 1, 2020 will contain only ARTFL & LEFFTDS data. Future data requests may select to include ARTFL and/or LEFFTDS data only.

• Acceptance of ALLFTD data obligates the recipient to cite/reference the GRANT in any presentation or publication that may result from this research. Language will be included in each ALLFTD publication following listed authors. Please see paragraph regarding acknowledgement below.

• Should publications result from the use of ALLFTD data now or in the future, the recipient agrees to notify the ALLFTD Executive Committee with details (reference or PubMedCentral ID#) and provide a copy of the publication so that the projects may report productivity derived from our resources to the funding agencies, the NIA, and NINDS. Such publications require compliance with National Institutes for Health (NIH) public access policies.

• Should funding result from this research now or in the future, please notify the ALLFTD Executive Committee (contact information below) with details (grant title, sponsor, number, dollar total, and dates) so that ALLFTD may report productivity derived from our resources to NIH.

• As described in the “Returning results” section above, new data created through analysis of ALLFTD data must be provided to the Executive Committee for possible inclusion in the ALLFTD database and other NIH‐approved governmental databases. Such data may be distributed in future ALLFTD data sets. Returning results also includes notification of negative findings, such that a manuscript will not be forthcoming.

• No sharing of data with a third party is allowed without permission of the ALLFTD Executive Committee.

For levels 1 and 2 Analyses:

“Data collection and dissemination of the data presented in this manuscript was supported by the ALLFTD Consortium (U19: AG063911, funded by the National Institute on Aging and the National Institute of Neurological Diseases and Stroke) and the former ARTFL & LEFFTDS Consortia (ARTFL: U54 NS092089, funded by the National Institute of Neurological Diseases and Stroke and National Center for Advancing Translational Sciences; LEFFTDS: U01 AG045390, funded by the National Institute on Aging and the National Institute of Neurological Diseases and Stroke). The manuscript has been reviewed by the ALLFTD Executive Committee for scientific content. The authors acknowledge the invaluable contributions of the study participants and families as well as the assistance of the support staffs at each of the participating sites.”

For Levels 3 Analyses:

“Data collection and dissemination of the data presented in this manuscript was supported by the ALLFTD Consortium (U19: AG063911, funded by the National Institute on Aging and the National Institute of Neurological Diseases and Stroke) and the former ARTFL & LEFFTDS Consortia (ARTFL: U54 NS092089, funded by the National Institute of Neurological Diseases and Stroke and National Center for Advancing Translational Sciences; LEFFTDS: U01 AG045390, funded by the National Institute on Aging and the National Institute of Neurological Diseases and Stroke). The authors acknowledge the invaluable contributions of the study participants and families as well as the assistance of the support staffs at each of the participating sites.”