The Control group protocol is identical to that of the Robotic group, except the robotically-incentivized exercise is replaced with an aerobic exercise. This intervention emulates a commercial motorized exercise bike to improve cardiovascular unfitness contributions to gait impairment. This therapy implements assist-as-needed and constant-velocity control. The patient's target pedaling speed (e.g. 20 rev/min (to be finalized through pilot tests) and power level (set by heart rate to require approximately 50-70% of maximal oxygen uptake) are set at the beginning of each session and a motor provides assistance or resistance to compensate for the performance of the patient. The exercise will be performed in this mode for 30 minutes per training session.

A simple test of the speed a person chooses to walk over a 5-meter distance

A similar test of the "fastest comfortable speed" a person can use to walk over the 5-meter distance

Movement of the Center of Pressure during standing, with eyes open and eyes closed. With eyes open, subjects focus their eyes on a spot on the wall in front of them. Metrics of COP control include excursion, average velocity, standard deviation, and an estimate of center of mass location.

Robotic therapy

Subjects with Stroke who will perform the targeted training task: exercise using novel tasks on a robotic recumbent cycle.

Aerobic therapy.

Subjects with Stroke who will perform aerobic pedaling, duration-matched to the Robotic group.

fMRI scans will be used to measure task-specific cortical activity in sensorimotor areas during physical stepping with an MRI-compatible stepping machine, or imagined/visualized cycling.

Healthy Pilot Group participants were enrolled and the study team learned that the protocol was insufficient to test the hypothesis. No data were collected for this measure.

A simple timed test of how long it takes to stand up, walk 10 feet, turn around, return to the starting point, and sit down.

Healthy Pilot Group participants were enrolled and the study team learned that the protocol was insufficient to test the hypothesis. No data were collected for this measure.

A simple test of the speed a person chooses to walk over a 5-meter distance

Healthy Pilot Group participants were enrolled and the study team learned that the protocol was insufficient to test the hypothesis. No data were collected for this measure.

A similar test of the "fastest comfortable speed" a person can use to walk over the 5-meter distance

Healthy Pilot Group participants were enrolled and the study team learned that the protocol was insufficient to test the hypothesis. No data were collected for this measure.

Movement of the Center of Pressure during standing, with eyes open and eyes closed. With eyes open, subjects focus their eyes on a spot on the wall in front of them. Metrics of COP control include excursion, average velocity, standard deviation, and an estimate of center of mass location.

Healthy Pilot Group participants were enrolled and the study team learned that the protocol was insufficient to test the hypothesis. No data were collected for this measure.

The test involved reaching against spring-like torque, where the spring stiffness was mostly constant but was changed unexpectedly by the computer in different trials. Reaction time definition: time from a visual and auditory cue to initial movement.

Participants were "Trained" with mostly a Low or High stiffness (each in a separate block of reaches) and had "Catch" trials when the stiffness was changed to Medium.

"Low" stiffness = 10 Newton meter per radian (Nm/rad); "Medium" stiffness = 20 Nm/rad; "High" stiffness = 30 Nm/rad

Original study plan required re-evaluation and the decision was made to pivot to developing tests and exercises on healthy population. A clinical trial was not conducted. Summary Data collected from experimental tasks in development are reported here as post-hoc outcomes.

The test involved reaching against spring-like torque, where the spring stiffness was mostly constant but was changed unexpectedly by the computer in different trials. Movement Settling Time definition: time for crank angle to settle between 40 and 50 degrees, when the central target was 45 degrees.

"Low" stiffness = 10 Nm/rad; "Medium" stiffness = 20 Nm/rad; "High" stiffness = 30 Nm/rad

"Catch trials" are unexpected medium stiffness setting, unpredictably spaced.

Original study plan required re-evaluation and the decision was made to pivot to developing tests and exercises on healthy population. A clinical trial was not conducted. Summary Data collected from experimental tasks in development are reported here as post-hoc outcomes.

The test involved reaching against spring-like torque, where the spring stiffness was mostly constant but was changed unexpectedly by the computer in different trials. All participants had a block with "Low" training, called "Train Low Catch Medium", and a separate block with "High" training, called "Train High Catch Medium".

Undershoot: initial reach showed undershoot relative to the target crank angle of 45 degrees, with the value measured as the local minimum as the pedal momentarily stopped and went backward.

Overshoot: initial reach showed overshoot beyond the target crank angle of 45 degrees, with the value measured as the maximum angle during that period.

Original study plan required re-evaluation and the decision was made to pivot to developing tests and exercises on healthy population. A clinical trial was not conducted. Summary Data collected from experimental tasks in development are reported here as post-hoc outcomes.

The test involved reaching against viscous damper torque, where the damping constant (units: Nm/(rad/s) = torque produced by the robot per rad/s of crank velocity) was changed by the computer in a structured but unpredictable sequence, across the trials. Movement settling time is defined as time for crank angle to settle between 40 and 50 degrees, when the central target was 45 degrees.

Low damping = 3 Nm/(rad/s), High damping = 9 Nm/(rad/s).

Original study plan required re-evaluation and the decision was made to pivot to developing tests and exercises on healthy population. A clinical trial was not conducted. Summary Data collected from experimental tasks in development are reported here as post-hoc outcomes.

The test involved reaching against viscous damper torque, where the damping constant (units: Nm/(rad/s) = torque produced by the robot per rad/s of crank velocity) was changed by the computer in a structured but unpredictable sequence, across the trials. Movement settling time is defined as time for crank angle to settle between 40 and 50 degrees, when the central target was 45 degrees. In this experiment, the damping level was visually cued. Low damping = 3 Nm/(rad/s), High damping = 9 Nm/(rad/s)

Original study plan required re-evaluation and the decision was made to pivot to developing tests and exercises on healthy population. A clinical trial was not conducted. Summary Data collected from experimental tasks in development are reported here as post-hoc outcomes.

The effect of pushing on one specific pedal was reversed by the computer in its calculation of how to move the crank; the other pedal had a physically typical effect. The task was to cause this complex system to move the crank at a constant 20 RPM forward speed (i.e., minimize the error in speed from this target). Reported here are the number of participants who completed this task.

Original study plan required re-evaluation and the decision was made to pivot to developing tests and exercises on healthy population. A clinical trial was not conducted. Summary Data collected from experimental tasks in development are reported here as post-hoc outcomes.