Change in executive function will be assessed by comparing Pre-Test, Mid-Test (~6 months), and Post-Test (~12 months) performance on Matrix Reasoning Accuracy from the Virginia Cognitive Aging Project (VCAP). Version O was given at pre, Version A was given at mid, and Version B was given at post. Possible scores range from 0-1 (i.e., 0% accuracy to 100% accuracy expressed on a 0.00-1.00 scale) with a higher score being indicative of better performance.

Change in memory will be assessed by comparing Pre-Test, Mid-Test (~6 months), and Post-Test (~12 months) performance on the Auditory Verbal Learning Test. Majdan et al. (1996) Form 1 was given at pre, the standard Rey Auditory Verbal Learning list was given at mid, and Majdan et al. (1996) Form 2 was given at post. These versions are comparable in terms of their scores (Sherman, Tan, & Hrabok, 2022). Possible scores range from 0-15 with a bigger score being indicative of better memory.

Change in memory will be assessed by comparing Pre-Test, Mid-Test (~6 months), and Post-Test (~12 months) performance on the Rey Osterreith Complex Figure Test. Tests were scored using the deep learning approach for automated scoring published by Langer et al. eLife 2024;13:RP96017. DOI: https://doi.org/10.7554/eLife.96017. Possible scores range from 0-36 with higher scores being indicative of better memory.

Change in memory will be assessed by comparing Pre-Test, Mid-Test (~6 months), and Post-Test (~12 months) performance on the NIH Toolbox Picture Sequence Memory Test - The computed score is an IRT-based theta score of the number of adjacent pairs placed for trial 1 and 2 with a higher score being indicative of better memory. This score is automatically created within the NIH Toolbox. Version A was given at pre, Version B was given at mid, and Version C was given at post. Possible scores range from 200-700 with a higher score being indicative of better performance.

Change in memory will be assessed by comparing Pre-Test, Mid-Test (~6 months), and Post-Test (~12 months) Lure Discimination Index (LDI) on the Mnemonic Similarity Test (MST). The MST was created based upon the publicly available and validated instructions and stimuli provided by the creators of the task (Stark, Kirwan, & Stark, 2019). Equivalent versions of the test are administered so participants see different stimuli at each testing session and those stimuli were counterbalanced to rotate through the different item types. The formula for calculating the LDI is: p(similar|lure) - p(similar|foil) where p(similar|lure) = Total # of "similar" responses to lure items / 36 and p(similar|foil) = Total # "similar" responses to foil items / 36. Scores range from -1 to +1 with a larger score indicating better discrimination performance.

Change in attention will be assessed by comparing Pre-Test, Mid-Test (~6 months), and Post-Test (~12 months) performance on the 2-sec trial of the Paced Auditory Serial Addition Task. Scores are reported as the number of correct responses with possible scores ranging from 0-60 and larger scores being indicative of better attention.

Change in attention will be assessed by comparing Pre-Test, Mid-Test (~6 months), and Post-Test (~12 months) performance on the Neurological Assessment Battery (NAB) Digit Span Forward test. Scores are reported as the number of correct trials with possible scores ranging from 0-14 and larger scores being indicative of better attention.

Change in working memory will be assessed by comparing Pre-Test, Mid-Test (~6 months), and Post-Test (~12 months) performance on the NIH Toolbox List Sort Working Memory Test. Scores are reported as the sum of the total number of items correctly recalled and sequenced on Lists 1 and 2 with possible scores ranging from 0-26 and larger scores being indicative of better working memory.

Change in working memory will be assessed by comparing Pre-Test, Mid-Test (~6 months), and Post-Test (~12 months) reaction time on the 4-dot trial of the Spatial Working Memory Test. Scores are reported as the average reaction time (RT) for correct responses. Smaller scores are indicative of faster performance (i.e., better performance).

Change in working memory will be assessed by comparing Pre-Test, Mid-Test (~6 months), and Post-Test (~12 months) accuracy on the Neurological Assessment Battery (NAB) Digit Span Backwards test. Scores are reported as the number of correct trial with possible scores ranging from 0-14 and larger scores being indicative of better working memory.

Change in working memory will be assessed by comparing Pre-Test, Mid-Test (~6 months), and Post-Test (~12 months) performance on the Digit Symbol Modalities Test (SDMT) (Smith, 1973). Scores are reported as the sum of Oral and Written Trial Raw Scores (correct responses) with possible scores ranging from 0-220 and larger scores being indicative of better working memory.

Change in processing speed will be assessed by comparing Pre-Test, Mid-Test (~6 months), and Post-Test (~12 months) performance on Stroop Word condition (also known as the Stroop Congruent condition). RT is recorded in msec and a larger score is indicative of slower performance (worse performance)

Change in processing speed will be assessed by comparing Pre-Test, Mid-Test (~6 months), and Post-Test (~12 months) performance on Stroop Color condition (also known as the Stroop Neutral condition). RT is recorded in msec and a larger score is indicative of slower performance (worse performance)

Change in executive function will be assessed by comparing Pre-Test, Mid-Test (~6 months), and Post-Test (~12 months) performance on the Comprehensive Trail Making Test (TMT) A (TMT A) (also known as Trail 1). Performance is recorded in seconds with a larger score indicative of slower performance (worse performance)

Change in Left Hippocampus Volume will be assessed by comparing Pre-Test, Mid-Test (~6 months), and Post-Test (~12 months) for total segmented left hippocampus volume (i.e., left hippocampus segmented away from other regions). This is measured as volume (mm3), ranges from 2500 to 4500, and higher values are better.

Change in Right Hippocampus Volume will be assessed by comparing Pre-Test, Mid-Test (~6 months), and Post-Test (~12 months) for total segmented right hippocampus volume (i.e., right hippocampus segmented away from other regions). This is measured as volume (mm3), ranges from 2500 to 4500, and higher values are better.

Blood samples will be taken following a 12-hour fast. Assays will be conducted for brain-derived neurotrophic factor (BDNF), irisin, insulin-like growth factor (IGF)-1, glucose, insulin, tumor necrosis factor (TNF)-⍺, serum amyloid protein (SAP), albumin, apolipoprotein E (ApoE) and ⍺-2 macroglobulin. Change from pre-test to post-test will be assessed.

Blood samples will be taken following a 12-hour fast. Assays will be conducted for brain-derived neurotrophic factor (BDNF), irisin, insulin-like growth factor (IGF)-1, glucose, insulin, tumor necrosis factor (TNF)-⍺, serum amyloid protein (SAP), albumin, apolipoprotein E (ApoE) and ⍺-2 macroglobulin. Change from pre-test to post-test will be assessed.

Change in blood biomarkers (BDNF, irisin, IGF-1, glucose, insulin, TNF-⍺, serum amyloid protein (SAP), albumin, ApoE and ⍺-2 macroglobulin) - IGF-1. Blood samples will be taken following a 12-hour fast. Assays will be conducted for brain-derived neurotrophic factor (BDNF), irisin, insulin-like growth factor (IGF)-1, glucose, insulin, tumor necrosis factor (TNF)-⍺, serum amyloid protein (SAP), albumin, apolipoprotein E (ApoE) and ⍺-2 macroglobulin. Change from pre-test to post-test will be assessed.

The capture antibodies in the kits we received from the company for IGF-1 were not effective. We attempted to secure new kits to conduct the assays, but did not receive the kits in a timely fashion. We intend to conduct these assays upon receipt of the kits and the availability of personnel to perform the assays. The anticipated reporting date for IGF-1 is January 2027.

Blood samples will be taken following a 12-hour fast. Assays will be conducted for brain-derived neurotrophic factor (BDNF), irisin, insulin-like growth factor (IGF)-1, glucose, insulin, tumor necrosis factor (TNF)-⍺, serum amyloid protein (SAP), albumin, apolipoprotein E (ApoE) and ⍺-2 macroglobulin. Change from pre-test to post-test will be assessed.

Blood samples will be taken following a 12-hour fast. Assays will be conducted for brain-derived neurotrophic factor (BDNF), irisin, insulin-like growth factor (IGF)-1, glucose, insulin, tumor necrosis factor (TNF)-⍺, serum amyloid protein (SAP), albumin, apolipoprotein E (ApoE) and ⍺-2 macroglobulin. Change from pre-test to post-test will be assessed.

Blood samples will be taken following a 12-hour fast. Assays will be conducted for brain-derived neurotrophic factor (BDNF), irisin, insulin-like growth factor (IGF)-1, glucose, insulin, tumor necrosis factor (TNF)-⍺, serum amyloid protein (SAP), albumin, apolipoprotein E (ApoE) and ⍺-2 macroglobulin. Change from pre-test to post-test will be assessed.

Blood samples will be taken following a 12-hour fast. Assays will be conducted for brain-derived neurotrophic factor (BDNF), irisin, insulin-like growth factor (IGF)-1, glucose, insulin, tumor necrosis factor (TNF)-⍺, serum amyloid protein (SAP), albumin, apolipoprotein E (ApoE) and ⍺-2 macroglobulin. Change from pre-test to post-test will be assessed.

Blood samples will be taken following a 12-hour fast. Assays will be conducted for brain-derived neurotrophic factor (BDNF), irisin, insulin-like growth factor (IGF)-1, glucose, insulin, tumor necrosis factor (TNF)-⍺, serum amyloid protein (SAP), albumin, apolipoprotein E (ApoE) and ⍺-2 macroglobulin. Change from pre-test to post-test will be assessed.

Blood samples will be taken following a 12-hour fast. Assays will be conducted for brain-derived neurotrophic factor (BDNF), irisin, insulin-like growth factor (IGF)-1, glucose, insulin, tumor necrosis factor (TNF)-⍺, serum amyloid protein (SAP), albumin, apolipoprotein E (ApoE) and ⍺-2 macroglobulin. Change from pre-test to post-test will be assessed.

Blood samples will be taken following a 12-hour fast. Assays will be conducted for brain-derived neurotrophic factor (BDNF), irisin, insulin-like growth factor (IGF)-1, glucose, insulin, tumor necrosis factor (TNF)-⍺, serum amyloid protein (SAP), albumin, apolipoprotein E (ApoE) and ⍺-2 macroglobulin. Change from pre-test to post-test will be assessed.

Fitness will be assessed by comparing predicted VO2max at Pre-Test, Mid-Test (~6 months), and Post-Test (~12 months). Fitness will be assessed using a submaximal graded aerobic exercise test performed on a treadmill. Oxygen uptake (VO2) at each stage is estimated based on the treadmill speed and grade during a ramped exercise protocol performed until volitional exhaustion or test termination due to symptom limitations. The VO2max estimation uses the slope of the regression line between the heart rates (HR) of the last two stages to extrapolate to the participant's predicted VO2 at their age-predicted max HR (220-age). If HR at either of the last two stages is <110, VO2 could not be reliably calculated, and we consider the participant to have insufficient data (ID). We have explored other options for estimating VO2 max, but believe this is the best option available.

Functional MRI will be used to measure brain activity including resting-state connectivity and change will be assessed from pre-test to post-test.

Functional MRI will be used to measure brain activity including resting-state connectivity and change will be assessed from pre-test to post-test.

Functional MRI will be used to measure brain activity including resting-state connectivity and change will be assessed from pre-test to post-test.

Functional MRI will be used to measure brain activity including resting-state connectivity and change will be assessed from pre-test to post-test.

Functional MRI will be used to measure brain activity including resting-state connectivity and change will be assessed from pre-test to post-test.

Functional MRI will be used to measure brain activity including resting-state connectivity and change will be assessed from pre-test to post-test.

Total of all reporting groups

Change in executive function will be assessed by comparing Pre-Test, Mid-Test (~6 months), and Post-Test (~12 months) performance on the Comprehensive Trail Making Test (TMT) interference. TMT interference is calculated as total time to complete TMT B (also called Trail 5) minus the total time to complete TMT A (also called Trail 1)which were both recorded in seconds. TMT interference is, therefore, recorded in seconds with a larger score indicative of greater interference (worse performance)

11 PAC, 10 UCC dropped from pre to mid; 11 PAC, 5 did not attend mid but remained in the study. 3 PAC, 1 UCC dropped from mid to post. 4 PAC not at mid dropped before post. 1 PAC, 4 UCC did not attend post. We lost data from 1 PAC and 1 UCC at pre and 1 UCC at post due to scoring errors. Thus, we report data for 97 PAC and 81 UCC at pre, 76 PAC and 67 UCC at mid, and 76 PAC and 66 UCC at post.

Change in executive function will be assessed by comparing Pre-Test, Mid-Test (~6 months), and Post-Test (~12 months) performance on the NIH Toolbox Dimentional Change Card Sort Test (DCCS). We used the recommended outcome of the DCCS computed score. This is calculated automatically within the NIH Toolbox based upon a two-stage process combining a score of 0-5 from the accuracy data with a score of 0-5 from the reaction time data. Thus, scores range from 0-10 with a higher score indicative of better performance.

11 PAC, 10 UCC dropped from pre to mid; 11 PAC, 5 did not attend mid but remained in the study. 3 PAC, 1 UCC dropped from mid to post. 4 PAC not at mid dropped before post. 1 PAC, 4 UCC did not attend post. We lost data from 2 UCC at mid and 3 PAC at post (computer error). We report data from 98 PAC and 82 UCC at pre, 76 PAC and 65 UCC at mid, and 73 PAC and 67 UCC at post.

Change in executive function will be assessed by comparing Pre-Test, Mid-Test (~6 months), and Post-Test (~12 months) performance on the NIH Toolbox Flanker Inhibitory Control and Attention Test. We used the recommended outcome of the Flanker computed score. This is calculated automatically within the NIH Toolbox based upon a two-stage process combining a score of 0-5 from the accuracy data with a score of 0-5 from the reaction time data. Thus, scores range from 0-10 with a higher score indicative of better performance.

11 PAC, 10 UCC dropped from pre to mid; 11 PAC, 5 did not attend mid but remained in the study. 3 PAC, 1 UCC dropped from mid to post. 4 PAC not at mid dropped before post. 1 PAC, 4 UCC did not attend post. We lost data from 2 UCC at mid and 4 PAC at post (computer error). We report data from 98 PAC and 82 UCC at pre, 76 PAC and 65 UCC at mid, and 72 PAC and 67 UCC at post.

Change in executive function will be assessed by comparing Pre-Test, Mid-Test (~6 months), and Post-Test (~12 months) performance on Matrix Reasoning Accuracy from the Virginia Cognitive Aging Project (VCAP). Version O was given at pre, Version A was given at mid, and Version B was given at post. Possible scores range from 0-1 (i.e., 0% accuracy to 100% accuracy expressed on a 0.00-1.00 scale) with a higher score being indicative of better performance.

11 PAC, 10 UCC dropped from pre to mid; 11 PAC, 5 did not attend mid but remained in the study. 3 PAC, 1 UCC dropped from mid to post. 4 PAC not at mid dropped before post. 1 PAC, 4 UCC did not attend post. We lost data from 1 UCC and 1 PAC at mid and 1 PAC at post (computer error). We report data from 98 PAC and 82 UCC at pre, 75 PAC and 66 UCC at mid, and 75 PAC and 67 UCC at post.

Change in memory will be assessed by comparing Pre-Test, Mid-Test (~6 months), and Post-Test (~12 months) performance on the Auditory Verbal Learning Test. Majdan et al. (1996) Form 1 was given at pre, the standard Rey Auditory Verbal Learning list was given at mid, and Majdan et al. (1996) Form 2 was given at post. These versions are comparable in terms of their scores (Sherman, Tan, & Hrabok, 2022). Possible scores range from 0-15 with a bigger score being indicative of better memory.

11 PAC, 10 UCC dropped from pre to mid; 11 PAC, 5 did not attend mid but remained in the study. 3 PAC, 1 UCC dropped from mid to post. 4 PAC not at mid dropped before post. 1 PAC, 4 UCC did not attend post. Thus, we report data from 98 PAC and 82 UCC at pre, 76 PAC and 67 UCC at mid, and 76 PAC and 67 UCC at post.

Change in memory will be assessed by comparing Pre-Test, Mid-Test (~6 months), and Post-Test (~12 months) performance on the Rey Osterreith Complex Figure Test. Tests were scored using the deep learning approach for automated scoring published by Langer et al. eLife 2024;13:RP96017. DOI: https://doi.org/10.7554/eLife.96017. Possible scores range from 0-36 with higher scores being indicative of better memory.

11 PAC, 10 UCC dropped from pre to mid; 11 PAC, 5 did not attend mid but remained in the study. 3 PAC, 1 UCC dropped from mid to post. 4 PAC not at mid dropped before post. 1 PAC, 4 UCC did not attend post. Thus, we report data from 98 PAC and 82 UCC at pre, 76 PAC and 67 UCC at mid, and 76 PAC and 67 UCC at post.

Change in memory will be assessed by comparing Pre-Test, Mid-Test (~6 months), and Post-Test (~12 months) performance on the NIH Toolbox Picture Sequence Memory Test - The computed score is an IRT-based theta score of the number of adjacent pairs placed for trial 1 and 2 with a higher score being indicative of better memory. This score is automatically created within the NIH Toolbox. Version A was given at pre, Version B was given at mid, and Version C was given at post. Possible scores range from 200-700 with a higher score being indicative of better performance.

11 PAC, 10 UCC dropped from pre to mid; 11 PAC, 5 did not attend mid but remained in the study. 3 PAC, 1 UCC dropped from mid to post. 4 PAC not at mid dropped before post. 1 PAC, 4 UCC did not attend post. We lost data from 2 UCC at mid and 3 PAC at post (computer error). We report data from 98 PAC and 82 UCC at pre, 76 PAC and 65 UCC at mid, and 73 PAC and 67 UCC at post.

Change in memory will be assessed by comparing Pre-Test, Mid-Test (~6 months), and Post-Test (~12 months) Lure Discimination Index (LDI) on the Mnemonic Similarity Test (MST). The MST was created based upon the publicly available and validated instructions and stimuli provided by the creators of the task (Stark, Kirwan, & Stark, 2019). Equivalent versions of the test are administered so participants see different stimuli at each testing session and those stimuli were counterbalanced to rotate through the different item types. The formula for calculating the LDI is: p(similar|lure) - p(similar|foil) where p(similar|lure) = Total # of "similar" responses to lure items / 36 and p(similar|foil) = Total # "similar" responses to foil items / 36. Scores range from -1 to +1 with a larger score indicating better discrimination performance.

11 PAC, 10 UCC dropped from pre to mid; 11 PAC, 5 did not attend mid but remained in the study. 3 PAC, 1 UCC dropped from mid to post. 4 PAC not at mid dropped before post. 1 PAC, 4 UCC did not attend post. We lost data from 5 UCC, 6 PAC at pre, from 4 UCC, 2 PAC at mid, and from 2 UCC, 2 PAC at post (computer error). We report data from 92 PAC, 77 UCC at pre, 74 PAC, 63 UCC at mid, and 75 PAC, 65 UCC at post.

Change in attention will be assessed by comparing Pre-Test, Mid-Test (~6 months), and Post-Test (~12 months) performance on the 2-sec trial of the Paced Auditory Serial Addition Task. Scores are reported as the number of correct responses with possible scores ranging from 0-60 and larger scores being indicative of better attention.

11 PAC, 10 UCC dropped from pre to mid; 11 PAC, 5 did not attend mid but remained in the study. 3 PAC, 1 UCC dropped from mid to post. 4 PAC not at mid dropped before post. 1 PAC, 4 UCC did not attend post. We lost data from 2 PAC at post (computer error). Thus, we report data from 98 PAC and 82 UCC at pre, 76 PAC and 67 UCC at mid, and 74 PAC and 67 UCC at post.

Change in attention will be assessed by comparing Pre-Test, Mid-Test (~6 months), and Post-Test (~12 months) performance on the Neurological Assessment Battery (NAB) Digit Span Forward test. Scores are reported as the number of correct trials with possible scores ranging from 0-14 and larger scores being indicative of better attention.

11 PAC, 10 UCC dropped from pre to mid; 11 PAC, 5 did not attend mid but remained in the study. 3 PAC, 1 UCC dropped from mid to post. 4 PAC not at mid dropped before post. 1 PAC, 4 UCC did not attend post. Thus, we report data from 98 PAC and 82 UCC at pre, 76 PAC and 67 UCC at mid, and 76 PAC and 67 UCC at post.

Change in working memory will be assessed by comparing Pre-Test, Mid-Test (~6 months), and Post-Test (~12 months) performance on the NIH Toolbox List Sort Working Memory Test. Scores are reported as the sum of the total number of items correctly recalled and sequenced on Lists 1 and 2 with possible scores ranging from 0-26 and larger scores being indicative of better working memory.

11 PAC, 10 UCC dropped from pre to mid; 11 PAC, 5 did not attend mid but remained in the study. 3 PAC, 1 UCC dropped from mid to post. 4 PAC not at mid dropped before post. 1 PAC, 4 UCC did not attend post. We lost data from 2 UCC at mid and 3 PAC at post (computer error). We report data from 98 PAC and 82 UCC at pre, 76 PAC and 65 UCC at mid, and 73 PAC and 67 UCC at post.

Change in working memory will be assessed by comparing Pre-Test, Mid-Test (~6 months), and Post-Test (~12 months) reaction time on the 4-dot trial of the Spatial Working Memory Test. Scores are reported as the average reaction time (RT) for correct responses. Smaller scores are indicative of faster performance (i.e., better performance).

11 PAC, 10 UCC dropped from pre to mid; 11 PAC, 5 did not attend mid but remained in the study. 3 PAC, 1 UCC dropped from mid to post. 4 PAC not at mid dropped before post. 1 PAC, 4 UCC did not attend post. We lost data from 1 UCC and 1 PAC at pre due to EPrime errors. Thus, we report data from 97 PAC and 81 UCC at pre, 76 PAC and 67 UCC at mid, and 76 PAC and 67 UCC at post.

Change in working memory will be assessed by comparing Pre-Test, Mid-Test (~6 months), and Post-Test (~12 months) accuracy on the Neurological Assessment Battery (NAB) Digit Span Backwards test. Scores are reported as the number of correct trial with possible scores ranging from 0-14 and larger scores being indicative of better working memory.

11 PAC, 10 UCC dropped from pre to mid; 11 PAC, 5 did not attend mid but remained in the study. 3 PAC, 1 UCC dropped from mid to post. 4 PAC not at mid dropped before post. 1 PAC, 4 UCC did not attend post. Thus, we report data from 98 PAC and 82 UCC at pre, 76 PAC and 67 UCC at mid, and 76 PAC and 67 UCC at post.

Change in working memory will be assessed by comparing Pre-Test, Mid-Test (~6 months), and Post-Test (~12 months) performance on the Digit Symbol Modalities Test (SDMT) (Smith, 1973). Scores are reported as the sum of Oral and Written Trial Raw Scores (correct responses) with possible scores ranging from 0-220 and larger scores being indicative of better working memory.

11 PAC, 10 UCC dropped from pre to mid; 11 PAC, 5 did not attend mid but remained in the study. 3 PAC, 1 UCC dropped from mid to post. 4 PAC not at mid dropped before post. 1 PAC, 4 UCC did not attend post. We lost one oral score from 1 UCC and 1 PAC at post. Thus, we report data from 98 PAC and 82 UCC at pre, 76 PAC and 67 UCC at mid, and 75 PAC and 66 UCC at post.

Change in processing speed will be assessed by comparing Pre-Test, Mid-Test (~6 months), and Post-Test (~12 months) performance on Stroop Word condition (also known as the Stroop Congruent condition). RT is recorded in msec and a larger score is indicative of slower performance (worse performance)

11 PAC, 10 UCC dropped from pre to mid; 11 PAC, 5 did not attend mid but remained in the study. 3 PAC, 1 UCC dropped from mid to post. 4 PAC not at mid dropped before post. 1 PAC, 4 UCC did not attend post. We lost data from 1 UCC participant at mid due to a computer error. Thus, we report data from 98 PAC and 82 UCC at pre, 76 PAC and 66 UCC at mid, and 76 PAC and 67 UCC at post.

Change in processing speed will be assessed by comparing Pre-Test, Mid-Test (~6 months), and Post-Test (~12 months) performance on Stroop Color condition (also known as the Stroop Neutral condition). RT is recorded in msec and a larger score is indicative of slower performance (worse performance)

11 PAC, 10 UCC dropped from pre to mid; 11 PAC, 5 did not attend mid but remained in the study. 3 PAC, 1 UCC dropped from mid to post. 4 PAC not at mid dropped before post. 1 PAC, 4 UCC did not attend post. We lost data from 1 UCC participant at mid due to a computer error. Thus, we report data from 98 PAC and 82 UCC at pre, 76 PAC and 66 UCC at mid, and 76 PAC and 67 UCC at post.

Change in executive function will be assessed by comparing Pre-Test, Mid-Test (~6 months), and Post-Test (~12 months) performance on the Comprehensive Trail Making Test (TMT) A (TMT A) (also known as Trail 1). Performance is recorded in seconds with a larger score indicative of slower performance (worse performance)

11 PAC, 10 UCC dropped from pre to mid; 11 PAC, 5 did not attend mid but remained in the study. 3 PAC, 1 UCC dropped from mid to post. 4 PAC not at mid dropped before post. 1 PAC, 4 UCC did not attend post. We lost data from 1 PAC and 1UCC at pre and 1 UCC at post due to scoring errors. Thus, we report data for 97 PAC and 81 UCC at pre, 76 PAC and 67 UCC at mid, and 76 PAC and 66 UCC at post.

Change in whole brain morphology will be assessed by comparing Pre-Test, Mid-Test (~6 months), and Post-Test (~12 months) for total segmented brain volume (i.e., total brain separated from non-brain structures). This is measured as volume (mm3), with volumes ranging from 1,050,000 to 1,350,000, and with higher values being interpreted as better.

We had 98 PAC and 82 UCC enrolled at pre. N=50 were not eligible for or unwilling to do the MRI (27 PAC, 24 UCC). At pre, 1 PAC could not complete her scan and there was missing data for 2 (1 PAC, 1 UCC) due to the pipeline used for preprocessing. 17 dropped before post (9 PAC, 8 UCC), 4 (2 PAC, 2 UCC) could not be scheduled, and there was missing data for 4 (2 PAC, 2 UCC) due to the pipeline. Data are therefore available for 70 PAC and 56 UCC at pre and 58 PAC and 46 UCC at post.

Change in Left Hippocampus Volume will be assessed by comparing Pre-Test, Mid-Test (~6 months), and Post-Test (~12 months) for total segmented left hippocampus volume (i.e., left hippocampus segmented away from other regions). This is measured as volume (mm3), ranges from 2500 to 4500, and higher values are better.

We had 98 PAC and 82 UCC enrolled at pre. N=50 were not eligible for or unwilling to do the MRI (27 PAC, 24 UCC). At pre, 1 PAC could not complete her scan and there was missing data for 2 (1 PAC, 1 UCC) due to the pipeline used for preprocessing. 17 dropped before post (9 PAC, 8 UCC), 4 (2 PAC, 2 UCC) could not be scheduled, and there was missing data for 4 (2 PAC, 2 UCC) due to the pipeline. Data are therefore available for 70 PAC and 56 UCC at pre and 58 PAC and 46 UCC at post.

Change in Right Hippocampus Volume will be assessed by comparing Pre-Test, Mid-Test (~6 months), and Post-Test (~12 months) for total segmented right hippocampus volume (i.e., right hippocampus segmented away from other regions). This is measured as volume (mm3), ranges from 2500 to 4500, and higher values are better.

We had 98 PAC and 82 UCC enrolled at pre. N=50 were not eligible for or unwilling to do the MRI (27 PAC, 24 UCC). At pre, 1 PAC could not complete her scan and there was missing data for 2 (1 PAC, 1 UCC) due to the pipeline used for preprocessing. 17 dropped before post (9 PAC, 8 UCC), 4 (2 PAC, 2 UCC) could not be scheduled, and there was missing data for 4 (2 PAC, 2 UCC) due to the pipeline. Data are therefore available for 70 PAC and 56 UCC at pre and 58 PAC and 46 UCC at post.

Blood samples will be taken following a 12-hour fast. Assays will be conducted for brain-derived neurotrophic factor (BDNF), irisin, insulin-like growth factor (IGF)-1, glucose, insulin, tumor necrosis factor (TNF)-⍺, serum amyloid protein (SAP), albumin, apolipoprotein E (ApoE) and ⍺-2 macroglobulin. Change from pre-test to post-test will be assessed.

Participants (N=14) dropped out from pre-test to mid-test (10 in PAC, 4 in UCC); Additional participants (N=17) did not attend the mid-test but remained in the study (n=10 in PAC, n=7 in UCC). Participants dropped out from mid-test to post-test (N=6, 5 in PAC, 1 in UCC); Additional participants (N=13) did not attend the post-test (n=5 in PAC, n=8 in UCC). Other missing data is because participants declined to provide a sample, we were unable to take a sample, or there was insufficient sample

Blood samples will be taken following a 12-hour fast. Assays will be conducted for brain-derived neurotrophic factor (BDNF), irisin, insulin-like growth factor (IGF)-1, glucose, insulin, tumor necrosis factor (TNF)-⍺, serum amyloid protein (SAP), albumin, apolipoprotein E (ApoE) and ⍺-2 macroglobulin. Change from pre-test to post-test will be assessed.

Participants (N=14) dropped out from pre-test to mid-test (10 in PAC, 4 in UCC); Additional participants (N=17) did not attend the mid-test but remained in the study (n=10 in PAC, n=7 in UCC). Participants dropped out from mid-test to post-test (N=6, 5 in PAC, 1 in UCC); Additional participants (N=13) did not attend the post-test (n=5 in PAC, n=8 in UCC). Other missing data is because participants declined to provide a sample, we were unable to take a sample, or there was insufficient sample

Change in blood biomarkers (BDNF, irisin, IGF-1, glucose, insulin, TNF-⍺, serum amyloid protein (SAP), albumin, ApoE and ⍺-2 macroglobulin) - IGF-1. Blood samples will be taken following a 12-hour fast. Assays will be conducted for brain-derived neurotrophic factor (BDNF), irisin, insulin-like growth factor (IGF)-1, glucose, insulin, tumor necrosis factor (TNF)-⍺, serum amyloid protein (SAP), albumin, apolipoprotein E (ApoE) and ⍺-2 macroglobulin. Change from pre-test to post-test will be assessed.

The capture antibodies in the kits we received from the company for IGF-1 were not effective. We attempted to secure new kits to conduct the assays, but did not receive the kits in a timely fashion. We intend to conduct these assays upon receipt of the kits and the availability of personnel to perform the assays. The anticipated reporting date for IGF-1 is January 2027.

Blood samples will be taken following a 12-hour fast. Assays will be conducted for brain-derived neurotrophic factor (BDNF), irisin, insulin-like growth factor (IGF)-1, glucose, insulin, tumor necrosis factor (TNF)-⍺, serum amyloid protein (SAP), albumin, apolipoprotein E (ApoE) and ⍺-2 macroglobulin. Change from pre-test to post-test will be assessed.

Participants (N=14) dropped out from pre-test to mid-test (10 in PAC, 4 in UCC); Additional participants (N=17) did not attend the mid-test but remained in the study (n=10 in PAC, n=7 in UCC). Participants dropped out from mid-test to post-test (N=6, 5 in PAC, 1 in UCC); Additional participants (N=13) did not attend the post-test (n=5 in PAC, n=8 in UCC). Other missing data is because participants declined to provide a sample, we were unable to take a sample, or there was insufficient sample

Blood samples will be taken following a 12-hour fast. Assays will be conducted for brain-derived neurotrophic factor (BDNF), irisin, insulin-like growth factor (IGF)-1, glucose, insulin, tumor necrosis factor (TNF)-⍺, serum amyloid protein (SAP), albumin, apolipoprotein E (ApoE) and ⍺-2 macroglobulin. Change from pre-test to post-test will be assessed.

Participants (N=14) dropped out from pre-test to mid-test (10 in PAC, 4 in UCC); Additional participants (N=17) did not attend the mid-test but remained in the study (n=10 in PAC, n=7 in UCC). Participants dropped out from mid-test to post-test (N=6, 5 in PAC, 1 in UCC); Additional participants (N=13) did not attend the post-test (n=5 in PAC, n=8 in UCC). Other missing data is because participants declined to provide a sample, we were unable to take a sample, or there was insufficient sample

Blood samples will be taken following a 12-hour fast. Assays will be conducted for brain-derived neurotrophic factor (BDNF), irisin, insulin-like growth factor (IGF)-1, glucose, insulin, tumor necrosis factor (TNF)-⍺, serum amyloid protein (SAP), albumin, apolipoprotein E (ApoE) and ⍺-2 macroglobulin. Change from pre-test to post-test will be assessed.

Participants (N=14) dropped out from pre-test to mid-test (10 in PAC, 4 in UCC); Additional participants (N=17) did not attend the mid-test but remained in the study (n=10 in PAC, n=7 in UCC). Participants dropped out from mid-test to post-test (N=6, 5 in PAC, 1 in UCC); Additional participants (N=13) did not attend the post-test (n=5 in PAC, n=8 in UCC). Other missing data is because participants declined to provide a sample, we were unable to take a sample, or there was insufficient sample

Blood samples will be taken following a 12-hour fast. Assays will be conducted for brain-derived neurotrophic factor (BDNF), irisin, insulin-like growth factor (IGF)-1, glucose, insulin, tumor necrosis factor (TNF)-⍺, serum amyloid protein (SAP), albumin, apolipoprotein E (ApoE) and ⍺-2 macroglobulin. Change from pre-test to post-test will be assessed.

Participants (N=14) dropped out from pre-test to mid-test (10 in PAC, 4 in UCC); Additional participants (N=17) did not attend the mid-test but remained in the study (n=10 in PAC, n=7 in UCC). Participants dropped out from mid-test to post-test (N=6, 5 in PAC, 1 in UCC); Additional participants (N=13) did not attend the post-test (n=5 in PAC, n=8 in UCC). Other missing data is because participants declined to provide a sample, we were unable to take a sample, or there was insufficient sample

Blood samples will be taken following a 12-hour fast. Assays will be conducted for brain-derived neurotrophic factor (BDNF), irisin, insulin-like growth factor (IGF)-1, glucose, insulin, tumor necrosis factor (TNF)-⍺, serum amyloid protein (SAP), albumin, apolipoprotein E (ApoE) and ⍺-2 macroglobulin. Change from pre-test to post-test will be assessed.

Participants (N=14) dropped out from pre-test to mid-test (10 in PAC, 4 in UCC); Additional participants (N=17) did not attend the mid-test but remained in the study (n=10 in PAC, n=7 in UCC). Participants dropped out from mid-test to post-test (N=6, 5 in PAC, 1 in UCC); Additional participants (N=13) did not attend the post-test (n=5 in PAC, n=8 in UCC). Other missing data is because participants declined to provide a sample, we were unable to take a sample, or there was insufficient sample

Blood samples will be taken following a 12-hour fast. Assays will be conducted for brain-derived neurotrophic factor (BDNF), irisin, insulin-like growth factor (IGF)-1, glucose, insulin, tumor necrosis factor (TNF)-⍺, serum amyloid protein (SAP), albumin, apolipoprotein E (ApoE) and ⍺-2 macroglobulin. Change from pre-test to post-test will be assessed.

Participants (N=14) dropped out from pre-test to mid-test (10 in PAC, 4 in UCC); Additional participants (N=17) did not attend the mid-test but remained in the study (n=10 in PAC, n=7 in UCC). Participants dropped out from mid-test to post-test (N=6, 5 in PAC, 1 in UCC); Additional participants (N=13) did not attend the post-test (n=5 in PAC, n=8 in UCC). Other missing data is because participants declined to provide a sample, we were unable to take a sample, or there was insufficient sample

Blood samples will be taken following a 12-hour fast. Assays will be conducted for brain-derived neurotrophic factor (BDNF), irisin, insulin-like growth factor (IGF)-1, glucose, insulin, tumor necrosis factor (TNF)-⍺, serum amyloid protein (SAP), albumin, apolipoprotein E (ApoE) and ⍺-2 macroglobulin. Change from pre-test to post-test will be assessed.

Participants (N=14) dropped out from pre-test to mid-test (10 in PAC, 4 in UCC); Additional participants (N=17) did not attend the mid-test but remained in the study (n=10 in PAC, n=7 in UCC). Participants dropped out from mid-test to post-test (N=6, 5 in PAC, 1 in UCC); Additional participants (N=13) did not attend the post-test (n=5 in PAC, n=8 in UCC). Other missing data is because participants declined to provide a sample, we were unable to take a sample, or there was insufficient sample

Fitness will be assessed by comparing predicted VO2max at Pre-Test, Mid-Test (~6 months), and Post-Test (~12 months). Fitness will be assessed using a submaximal graded aerobic exercise test performed on a treadmill. Oxygen uptake (VO2) at each stage is estimated based on the treadmill speed and grade during a ramped exercise protocol performed until volitional exhaustion or test termination due to symptom limitations. The VO2max estimation uses the slope of the regression line between the heart rates (HR) of the last two stages to extrapolate to the participant's predicted VO2 at their age-predicted max HR (220-age). If HR at either of the last two stages is <110, VO2 could not be reliably calculated, and we consider the participant to have insufficient data (ID). We have explored other options for estimating VO2 max, but believe this is the best option available.

At pre, we had insufficient data (ID) for 15 PAC, 16 UCC. 6 PAC, 4 UCC were not tested. 11 PAC, 10 UCC dropped from pre to mid; 11 PAC, 5 UCC did not attend mid. At mid, ID for 17 PAC, 5 UCC. 9 PAC, 14 UCC were not tested. 3 PAC, 1 UCC dropped from mid to post. 4 PAC not at mid dropped before post. 1 PAC, 4 UCC did not attend post. At post, ID for 15 PAC, 13 UCC. 6 PAC, 8 UCC were not tested. Thus, we report data from 77 PAC, 62 UCC at pre, 50 PAC, 48 UCC at mid, and 55 PAC, 46 UCC at post.

Functional MRI will be used to measure brain activity including resting-state connectivity and change will be assessed from pre-test to post-test.

N=50 were not eligible for or unwilling to do the MRI at pre or post (27 PAC, 23 UCC). At pre, 1 PAC could not be scheduled, and there was missing data for 4 (3 PAC, 1 UCC). Thus, data is available for 67 PAC and 58 UCC at pre. 17 dropped before post (10 PAC, 7 UCC), 2 PAC refused MRI, 3 UCC could not be scheduled, and there was missing data for 2 (1 PAC, 1 UCC). Thus, data is available for 59 PAC and 47 UCC at post.

Functional MRI will be used to measure brain activity including resting-state connectivity and change will be assessed from pre-test to post-test.

N=50 were not eligible for or unwilling to do the MRI at pre or post (27 PAC, 23 UCC). At pre, 1 PAC could not be scheduled, and there was missing data for 4 (3 PAC, 1 UCC). Thus, data is available for 67 PAC and 58 UCC at pre. 17 dropped before post (10 PAC, 7 UCC), 2 PAC refused MRI, 3 UCC could not be scheduled, and there was missing data for 2 (1 PAC, 1 UCC). Thus, data is available for 59 PAC and 47 UCC at post.

Functional MRI will be used to measure brain activity including resting-state connectivity and change will be assessed from pre-test to post-test.

N=50 were not eligible for or unwilling to do the MRI at pre or post (27 PAC, 23 UCC). At pre, 1 PAC could not be scheduled, and there was missing data for 4 (3 PAC, 1 UCC). Thus, data is available for 67 PAC and 58 UCC at pre. 17 dropped before post (10 PAC, 7 UCC), 2 PAC refused MRI, 3 UCC could not be scheduled, and there was missing data for 2 (1 PAC, 1 UCC). Thus, data is available for 59 PAC and 47 UCC at post.

Functional MRI will be used to measure brain activity including resting-state connectivity and change will be assessed from pre-test to post-test.

N=50 were not eligible for or unwilling to do the MRI at pre or post (27 PAC, 23 UCC). At pre, 1 PAC could not be scheduled, and there was missing data for 4 (3 PAC, 1 UCC). Thus, data is available for 67 PAC and 58 UCC at pre. 17 dropped before post (10 PAC, 7 UCC), 2 PAC refused MRI, 3 UCC could not be scheduled, and there was missing data for 2 (1 PAC, 1 UCC). Thus, data is available for 59 PAC and 47 UCC at post.

Functional MRI will be used to measure brain activity including resting-state connectivity and change will be assessed from pre-test to post-test.

N=50 were not eligible for or unwilling to do the MRI at pre or post (27 PAC, 23 UCC). At pre, 1 PAC could not be scheduled, and there was missing data for 4 (3 PAC, 1 UCC). Thus, data is available for 67 PAC and 58 UCC at pre. 17 dropped before post (10 PAC, 7 UCC), 2 PAC refused MRI, 3 UCC could not be scheduled, and there was missing data for 2 (1 PAC, 1 UCC). Thus, data is available for 59 PAC and 47 UCC at post.

Functional MRI will be used to measure brain activity including resting-state connectivity and change will be assessed from pre-test to post-test.

N=50 were not eligible for or unwilling to do the MRI at pre or post (27 PAC, 23 UCC). At pre, 1 PAC could not be scheduled, and there was missing data for 4 (3 PAC, 1 UCC). Thus, data is available for 67 PAC and 58 UCC at pre. 17 dropped before post (10 PAC, 7 UCC), 2 PAC refused MRI, 3 UCC could not be scheduled, and there was missing data for 2 (1 PAC, 1 UCC). Thus, data is available for 59 PAC and 47 UCC at post.

Participants in the usual care control will maintain their normal health practices for 1 year. Participants will receive a bi-weekly health newsletter and will be contacted bi-weekly to answer any questions and inquire about the participant's health. Participants self-reported physical activity will be assessed monthly. In this fashion, participants will be contacted by staff every week. Usual care control participants that complete all study related activities including pre-, mid-, and post-test will receive a short-term YMCA membership after post-test.