In many everyday activities such as steering a bicycle, unpredictable mechanical disturbances trigger quick involuntary and situation-specific reactions via neural loops generally referred to as reflexes. Research investigating reflexes via transient perturbations elicit muscle activations typically observed as two distinct responses in an electromyography with the short-latency response M1 viewed as stereotyped and task-independent and the long-latency response M2 seen as task-dependent. In contrast, task-dependency of short-latency pathways (specifically velocity and force feedback) was demonstrated by studies using continuous perturbations and system identification techniques to separate reflex from voluntary contributions. This study addressed the opposing experimental findings by isolating reflex contributions in the human wrist joint using both the transient and the continuous approach simultaneously in conditions where reflex modulation is expected. Subjects (n = 11) held a manipulator handle which applied mechanical perturbations and imposed a virtual mechanical environment to the wrist. Increasing damping of the environment and reduced continuous perturbation bandwidth in a "maintain position" task (PT) decreased the subject’s mechanical joint admittance, increased excitatory velocity and force feedback obtained from a neuromuscular model, increased M2 but did not affect M1. Instructing subjects to "maintain force" increased the joint admittance, decreased M2 and counterproductive to the task increased M1 with respect to the PT. Results from this study indicate that the continuous approach does not condition M1 and that the reflexive feedback from short-latency pathways obtained via the presented neuromuscular model does not directly map to the M1 elicited by transient perturbations.