Hamburger (1980) has reported that tactile point localisation is more accurate in the medio-lateral (M-L) axis than in the proximo-distal (P-D) axis of the human forearm. The explanation of this observation is uncertain, however, due to a potential sensory-motor confound that was inherent to the methodology, which involved the subject marking the perceived stimulus site using an active movement of the contralateral hand. We have re-examined the question using a response mode that did not rely on the subject’s motor precision.

Twenty-eight (16 females, 12 males, mean ± S.D. age 21.6 ± 1.1 years), right-handed, healthy subjects participated with local ethics committee approval. A 13 X 13 point (5 mm separation) grid was marked in ink on the skin of the dorsal surface of the subject’s left forearm and the grid axes were labelled A-M (transverse) and 0-12 (longitudinal). A brief tactile stimulus was applied, using a von Frey hair (rating 150 mN) to each point of the central 7 X 7 section of the grid, in randomised order, whilst the subject’s eyes were closed. Following each stimulus, the subject opened his eyes and stated the grid reference corresponding to the perceived stimulus locus. At each stimulus site, locognostic errors in the M-L and P-D axes were calculated by comparison of the stated grid reference with that of the stimulus locus. Two arm positions were tested, namely ‘straight arm’ (elbow extended) and ‘bent arm’ (elbow 90 deg).

For the straight arm position, the mean (S.D.) locognostic errors in the P-D and M-L axes were, respectively, 12.1 (4.3) mm and 8.0 (2.3) mm. The corresponding errors for the bent arm position were 11.5 (2.5) mm and 8.8 (3.2) mm. A 2-way, repeated-measures ANOVA indicated a significant main effect of axis (F (1, 27) = 39.78; P < 0.0005), whilst neither the effect of arm position (F (1, 27) = 0.01; P = 0.94) nor the axis-position interaction (F (1, 27) = 3.69; P = 0.07) was significant. Post-hoc paired t tests demonstrated that locognostic errors in the P-D axis exceeded those in the M-L axis for both the straight arm (t = 6.47, P < 0.005) and bent arm (t = 3.87; P = 0.04) positions.

Our finding of superior locognostic acuity in the transverse axis of the forearm, which confirms earlier reports using different methods, probably arises principally from the intrinsic innervation patterns of the limb, reflected in the oval-shaped, longitudinally orientated receptive fields of local sensory units (e.g. Schady et al. 1983).