A functionally meaningful vestibular-neck interaction, such as it has been demonstrated for postural reflexes and self-motion perception, requires the spatial and temporal response characteristics of vestibular and neck signals to be similar. We investigated the spatial coding in neurons of the external cuneate nucleus (ECN) with natural neck and vestibular stimulations, and compared them to that of neurons in the descending and medial vestibular nuclei (DVN and MVN, respectively) obtained with vestibular stimulation. Neurons were recorded extracellularly in chronically prepared cats held under light barbiturate anesthesia. Neck stimulation was performed by sinusoidally rotating the animals' trunk relative to the earth-fixed head in six different vertical planes and in the horizontal plane. Vestibular stimulation was elicited by whole-body rotations in the corresponding planes. During neck stimulation in the vertical planes, most ECN neurons showed an approximately sinusoidal discharge modulation about resting rate, which became maximal during rotation in a specific plane. Off this plane, the response declined along a cosine function and reached zero in the orthogonal plane. The majority of these ECN neurons also responded to horizontal neck rotation; the resulting "optimal" direction of rotation in three-dimensional space varied considerably among the neurons. Yet, there was a certain preference; the majority of these ECN neurons fired maximally if trunk rotation in the yaw plane stretched the neck on the ipsilateral side, if roll brought the contralateral shoulder closer to the head, and if pitch brought the back closer to the occiput. A minority of ECN neurons showed more complex response patterns which could not be described by a single, optimal direction. About one third of the neck-sensitive ECN neurons tested showed weak responses during whole body rotation, which might stem from a weak vestibular input to this nucleus. In the DVN and MVN, the optimal direction in three-dimensional space with vestibular stimulation typically had a cosine-like spatial tuning. The spatial distribution of these directions clearly differed from that of neck-sensitive neurons in the ECN. We therefore assume that a further processing of the two input signals takes place at later stages in the CNS (e.g., in the vestibulo-cerebellum) in order to yield a functionally useful vestibular-neck interaction.