Participants in this group will continue to receive their usual care. They will also receive activity monitors to monitor their daily activity levels, and telephone monitoring throughout the study. Study team members telephone all participants every 2 weeks for similar amounts of contact time between treatment groups. During the calls, a study team member of the reviews fall diary information, inquires about possible adverse events, and reviews participants' activity monitor wear times via the Garmin Connect website (https://connect.garmin.com).

The 10 Meter Walk Test (10MWT) evaluates gait speed similarly in separate trials with the participant instructed to walk at comfortable and fast pace, and velocity timed over the middle six meters. This outcome is the velocity (meters/second) with instructions to walk at a fast pace, performed off medication. Higher velocities indicate better motor function.

iPad-facilitated Timed Up and Go Test duration, measured off medication. The Timed "Up & Go" (TUG) Test requires an individual to (1) stand from a chair, (2) ambulate 3 m, (3) turn 180 degrees, (4) ambulate back to the chair, and (5) turn and sit. To perform an instrumented TUG, an Apple iPad is attached to the individual's waist (sacral level) via an adjustable belt. This outcome is the total time to complete this test, in seconds, measured off medication. Lower times indicate better motor function.

iPad-facilitated Timed Up and Go Test mean turning velocity, measured off medication. The Timed "Up & Go" (TUG) Test requires an individual to (1) stand from a chair, (2) ambulate 3 m, (3) turn 180 degrees, (4) ambulate back to the chair, and (5) turn and sit. To perform an instrumented TUG, an Apple iPad is attached to the individual's waist (sacral level) via an adjustable belt. This outcome is the mean velocity during performance of step (3), turn 180 degrees, in degrees/second, measured off medication. Higher turn velocities indicate better motor function.

Manual Dexterity Test time to completion in seconds, measured off medication. This test is an electronic (iPad-based), modified version of the 9-hole peg test, where an individual is required to grasp, transfer, and release pegs into a 3×3 grid. Participants are asked to grasp and place pegs into holes, and then to return them to their starting locations, as rapidly as possible. Completion times for two repetitions conducted with each hand were averaged. Shorter completion times indicate better motor function.

The Processing Speed Test is a self-administered electronic (iPad) facilitated symbol-number matching task that is similar to the Symbol Digital Modalities Test. Participants are shown a table with numerical labels for 9 disparate symbols, and a second table with a row of such symbols and a blank row beneath it, and asked to complete the second table with the numbers from the first table corresponding to each symbol in the second. When the table is completed, additional tables are shown. The response recorded is the number of correct matches of symbols with numbers in 120 seconds, measured off medication. Higher values indicate faster neural processing.

Participants completed iPad versions of Trail Making Tests A and B (TMT A and B, respectively), with durations in seconds recorded for each. TMT A, considered predominantly a motor task, requires the participant to use a stylus to connect dots in ascending numeric order (1-2-3-4 . . . ). The TMT B, requires the participant to connect dots in an ascending and alternating alpha and numeric order (1-A-2-B . . . .). The total times to complete each of the TMT A and B tasks are automatically recorded by the iPad. This outcome is the (unitless) ratio of the duration to complete Trail Making Test B to the duration to complete Trail Making Test A, reflecting the cognitive burden of visual attention and set-switching, measured off medication. Higher values reflect greater relative burden of the additional cognitive component.

Electronics-based test of episodic memory performed on an iPad, reported as number of correctly recalled symbols arrayed on a grid, with values ranging from 0 to 70 and higher numbers reflective of better memory, measured off medication.

Mild to moderate Parkinson's disease patients receiving no active exercise intervention through the research. These participants continued to receive usual and customary care (UCC) for their Parkinson's disease during the 12 month follow-up period. Participants receive and are asked to wear activity monitors throughout the day to monitor daily activity levels, and not to commence a cycling exercise program for the duration of the 12 month follow-up period. Participants also receive biweekly telephone calls to review electronic activity monitoring data, monitor falls and adverse events, and to provide them similar contact time as for the active exercise group.

Total of all reporting groups

Movement Disorder Society Unified Parkinson's Disease Rating Scale Part III: Motor Examination score, measured off medication. Range of scores is 0 to 132, with lower score indicating better motor function.

All observed data are summarized; 5 participants, 3 in the aerobic exercise (AE) and 2 in the usual care (UC) arm, withdrew prior to initiating randomized treatment and follow-up, of whom 4 had not yet completed the baseline MDS-UPDRS III. One participant was withdrawn by the study PI due to periodic atrial fibrillation, 2 withdrew due to COVID-19 pandemic-related logistical issues and safety concerns, and 2 withdrew due to dissatisfaction with their randomized usual care assignments.

The 10 Meter Walk Test (10MWT) evaluates gait speed similarly in separate trials with the participant instructed to walk at comfortable and fast pace, and velocity timed over the middle six meters. This outcome is the velocity (meters/second) with instructions to walk at a comfortable pace, performed off medication. Higher velocities indicate better motor function.

Medians and interquartile ranges are for the observed data, without imputation of missing data due to baseline or subsequent dropouts. Analytic results reported below are based on the full intention-to-treat population, with missing data multiply-imputed.

The 10 Meter Walk Test (10MWT) evaluates gait speed similarly in separate trials with the participant instructed to walk at comfortable and fast pace, and velocity timed over the middle six meters. This outcome is the velocity (meters/second) with instructions to walk at a fast pace, performed off medication. Higher velocities indicate better motor function.

Means and standard deviations are for the observed data, without imputation of missing data due to baseline or subsequent dropouts. Analytic results reported below are based on the full intention-to-treat population, with missing data multiply-imputed.

iPad-facilitated Timed Up and Go Test duration, measured off medication. The Timed "Up & Go" (TUG) Test requires an individual to (1) stand from a chair, (2) ambulate 3 m, (3) turn 180 degrees, (4) ambulate back to the chair, and (5) turn and sit. To perform an instrumented TUG, an Apple iPad is attached to the individual's waist (sacral level) via an adjustable belt. This outcome is the total time to complete this test, in seconds, measured off medication. Lower times indicate better motor function.

Medians and interquartile ranges are reported for the observed data, without imputation of missing data due to baseline or subsequent dropouts. Analytic modeling results reported below are based on the full intention-to-treat population, with missing data multiply-imputed.

iPad-facilitated Timed Up and Go Test mean turning velocity, measured off medication. The Timed "Up & Go" (TUG) Test requires an individual to (1) stand from a chair, (2) ambulate 3 m, (3) turn 180 degrees, (4) ambulate back to the chair, and (5) turn and sit. To perform an instrumented TUG, an Apple iPad is attached to the individual's waist (sacral level) via an adjustable belt. This outcome is the mean velocity during performance of step (3), turn 180 degrees, in degrees/second, measured off medication. Higher turn velocities indicate better motor function.

Means and standard deviations are for the observed data, without imputation of missing data due to baseline or subsequent dropouts. Analytic results reported below are based on the full intention-to-treat population, with missing data multiply-imputed.

Manual Dexterity Test time to completion in seconds, measured off medication. This test is an electronic (iPad-based), modified version of the 9-hole peg test, where an individual is required to grasp, transfer, and release pegs into a 3×3 grid. Participants are asked to grasp and place pegs into holes, and then to return them to their starting locations, as rapidly as possible. Completion times for two repetitions conducted with each hand were averaged. Shorter completion times indicate better motor function.

Medians and interquartile ranges are for the observed data, without imputation of missing data due to baseline or subsequent dropouts. Analytic results reported below are based on the full intention-to-treat population, with missing data multiply-imputed.

The Processing Speed Test is a self-administered electronic (iPad) facilitated symbol-number matching task that is similar to the Symbol Digital Modalities Test. Participants are shown a table with numerical labels for 9 disparate symbols, and a second table with a row of such symbols and a blank row beneath it, and asked to complete the second table with the numbers from the first table corresponding to each symbol in the second. When the table is completed, additional tables are shown. The response recorded is the number of correct matches of symbols with numbers in 120 seconds, measured off medication. Higher values indicate faster neural processing.

Medians and interquartile ranges are for the observed data, without imputation of missing data due to baseline or subsequent dropouts. Analytic results reported below are based on the full intention-to-treat population, with missing data multiply-imputed.

Participants completed iPad versions of Trail Making Tests A and B (TMT A and B, respectively), with durations in seconds recorded for each. TMT A, considered predominantly a motor task, requires the participant to use a stylus to connect dots in ascending numeric order (1-2-3-4 . . . ). The TMT B, requires the participant to connect dots in an ascending and alternating alpha and numeric order (1-A-2-B . . . .). The total times to complete each of the TMT A and B tasks are automatically recorded by the iPad. This outcome is the (unitless) ratio of the duration to complete Trail Making Test B to the duration to complete Trail Making Test A, reflecting the cognitive burden of visual attention and set-switching, measured off medication. Higher values reflect greater relative burden of the additional cognitive component.

Medians and interquartile ranges are for the observed data, without imputation of missing data due to baseline or subsequent dropouts. Analytic results reported below are based on the full intention-to-treat population, with missing data multiply-imputed.

Electronics-based test of episodic memory performed on an iPad, reported as number of correctly recalled symbols arrayed on a grid, with values ranging from 0 to 70 and higher numbers reflective of better memory, measured off medication.

Medians and quartiles are for the observed data, without imputation of missing data due to baseline or subsequent dropouts. Analytic modeling results reported below are based on the full intention-to-treat population, with missing data multiply-imputed.

Continue regular activity. No aerobic exercise intervention.