Text this: Participant followup rate can bias structural imaging measures in longitudinal studies