PACEMAKER ACTIVITY IN A SENSORY ENDING WITH MULTIPLE ENCODING SITES -THE CAT MUSCLE-SPINDLE PRIMARY ENDING

1. A combined physiological, histological and computer modelling study was carried out on muscle spindles of the cat tenuissimus muscle to e xamine whether there was any correlation between the functional intera ction of putative encoding sites, operated separately by static and dy namic fusimotor neurones, and the topological structure of the preterm inal branches of the primary sensory ending. 2. Spindles, whose Ia res ponses to stretch and separate and combined static and dynamic fusimot or stimulation were recorded in physiological experiments, were locate d in situ. Subsequently the ramifications of the sensory ending were r econstructed histologically, and the topology of the branch tree was u sed in computer simulations of I a responses to examine the effect of the electrotonic separation of encoding sites on the static-dynamic in teraction pattern. 3. Interactions between separate static and dynamic inputs, manifest in responses to combined stimulation, were quantifie d by a coefficient of interaction (C-1) which, by definition, was 1 fo r strictly linear summation of separate inputs and zero for maximum oc clusion between inputs. 4. For the majority of spindles static-dynamic interactions were characterized by pronounced occlusion (C-1 < 0.35). In these spindles putative encoding sites (the peripheral heminodes o f the branches supplying the intrafusal fibres activated by individual fusimotor efferents) were separated by a minimum conduction path of b etween three and ten myelinated segments (2-9 nodes of Ranvier). In co ntrast, significant summation (C-1, similar to 0.7) was found in only one spindle. In this case putative encoding sites were separated by a single node. 5. Occlusion was not due to encoder saturation and it cou ld not be accounted for by any other known physiological mechanisms (i ntrafusal fatigue or unloading). It is therefore attributed to competi tive pacemaker interaction between encoding sites which are largely se lectively operated by static and dynamic fusimotor efferents. 6. Model simulations of real preterminal-branch tree structures confirmed that short conduction paths between encoding sites were associated with ma nifest summation, whereas longer minimum conduction paths favoured pro nounced occlusion. 7. In the extreme, occlusion could be so pronounced as to give rise to negative values of C-1 during critical segments of response cycles. This was associated with lower discharge rates durin g combined static and dynamic stimulation than the higher of the indiv idual stimulation effects. This phenomenon is referred to as hyperoccl usion. Computer simulations demonstrated that hyperocclusion could be accounted for by a slow ionic adaptation process, e.g. by a very slowl y activating K+ conductance.