BACKGROUND Non-invasive brain stimulation techniques may be useful adjuvants to promote recovery after stroke. They are typically used to facilitate ipsilesional cortical excitability directly, or indirectly by suppressing contralesional cortical excitability and reducing interhemispheric inhibition from the contralesional to ipsilesional hemisphere. However, most of the evidence for this approach comes from studies of patients at the chronic stage of recovery. HYPOTHESIS We hypothesized that corticomotor excitability and interhemispheric inhibition would initially be asymmetric, with greater interhemispheric inhibition from contralesional to ipsilesional M1. We also hypothesized that balancing of corticomotor excitability and interhemispheric inhibition would be associated with greater improvements in paretic upper-limb impairment and function. METHODS We conducted a retrospective analysis of longitudinal data collected from 46 patients during the first six months after stroke. Transcranial magnetic stimulation was used to measure rest motor threshold, stimulus-response curves, and ipsilateral silent periods from the extensor carpi radialis muscles of both upper limbs. Analyses of variance and linear regression modeling were used to evaluate the effect of time on corticomotor excitability and interhemispheric inhibition in both hemispheres, and associations between these effects and improvements in paretic upper-limb impairment and function. RESULTS All participants had subcortical damage and only two had motor cortex involvement. As expected, ipsilesional corticomotor excitability was initially suppressed and increased over time, and this increase was associated with improved upper-limb impairment and function. However, interhemispheric inhibition was symmetrical and stable over time, and there was no evidence for a decrease in contralesional corticomotor excitability. CONCLUSIONS Neuromodulation interventions applied during spontaneous recovery may be more beneficial if they facilitate ipsilesional corticomotor excitability directly.