1036 IEEE TRANSACTIONS ON BIOMEDICAL ENGINEERING, VOL. 45, NO. 8, AUGUST 1998 Effects of Regional Stimulation Using a Miniature Stimulator Implanted in Feline Posterior Biceps Femoris Tracy Cameron,* Member, IEEE, Frances J. R. Richmond, and Gerald E. Loeb Abstract-The effects of placement of a miniature implantable that can be inserted percutaneously into muscles [7], [S], or stimulator on motor unit recruitment were examined in the posterior head of cat biceps femoris. The implantable stimulator (13-mm long x 2-mm diameter) was injected either proximally near the main nerve branch, or distally near the muscle inser- tion, through a 12-gauge hypodermic needle. Glycogen-depletion methods were used to map the distribution of fibers activated by electrical stimulation. Muscle fibers were found to be depleted at most or all proximodistal levels of the muscle, but the density of depleted fibers varied transversely according to the stimulus strength and proximity of the device to the nerve-entry site. Thus, muscle cross sections often had a "patchy" appearance produced because different proportions of depleted fibers intermingled with undepleted fibers in different parts of the cross section. In other preparations, the force of muscle contraction was measured when stimuli of varying strengths were delivered by the stimulator positioned at the same proximal or distal sites within the muscle. Devices placed close to the nerve-entry site produced the greatest forces. Those placed more distally produced less force. As stimu- lus current and/or pulse width increased, muscle force increased, often in steps, until a maximum was reached, which was usually limited by the compliance voltage of the device to less than the force produced by whole nerve stimdation. Index Tenns-Electrical stimulation, FES, FNS, paralysis, TES. N EURAL prostheses to reanimate muscle use electrical stimulation to activate intact nerves that have lost their natural inputs as a result of disease or injury to the central nervous system. The electrical stimuli cause contractions by activating the most excitable elements of the neuromuscular complex, the preterminal axons or terminal arborizations of myelinated motor nerves. Repeated, chronic stimulation of motor nerves supplying a paralyzed muscle can increase muscle strength and aerobic capacity [I]-[3], and produce functionally useful movements (for review see [4]). Several different approaches permit electrical stimulation of single muscles. These include the use of radio-frequency (RF) controlled, miniature single-channel stimulators that can be injected intramuscularly [5], [6], flexible wire electrodes Manuscript received December 20, 1996; revised February 24, 1998. This work was supported by the Ontario Rehabilitation Technology Consortium, the Canadian Neuroscience Network of Centres of Excellence, and the National Institutes of Health (NIH) under Contracts N01-NS-2-2322 and NOl-NS-5- 2325. Asterisk indicates corresponding author. *T. Cameron is with Advanced Neuromodulation Systems, Inc., 1 Allen- town Parkway, Allen, Texas 75002 USA (e-mail: [email protected]). F. J. R. Richmond and G. E. Loeb are with Queen's University, Kingston, Ont. K7L 3N6 Canada. Publisher Item Identifier S 0018-9294(98)05326-9. surgically implanted, RF-controlled multichannel stimulators whose wire electrodes are attached either to the surface of the muscle (epimysial) [9] or around the muscle nerve (epineural) [lo], [Ill. Epineural, intramuscular, and epimuscular electrodes are all known to produce muscle contractions whose force can be graded by increasing the strength of stimulation pulses. However, when stimulating electrodes are placed directly on nerves, it can be difficult to grade muscle force in small incre- ments because all of the alpha-motoneuron axons supplying different motor units have similar activation thresholds. Thus, small changes in stimulus strength can produce large shifts in motor-unit recruitment. Such shifts may occur even without changing the stimulus strength if the epineural electrodes move slightly with respect to the underlying axons in the nerve bundle. Embedded intramuscular electrodes are less sensitive to such problems because the greater intramuscular separation of motor nerve branches ensures a more gradual recruit- ment of motor axons when stimulus strength is increased. This advantage might have a "downside," however, when intramuscular devices are implanted in large muscles with specialized patterns of innervation and nonuniform motor-unit distribution. In such muscles, Merent parts of the muscle may be recruited nonunifonnly. Such nonunifonnities would introduce a significant level of complexity for a controller designed to produce a consistent level of muscle force for repeatable functional movements of a body part. In the present study, we have explored the hypothesis that the intramuscular locus of muscle stimulation has a significant effect upon the distribution of active motor units and the strength of muscle contraction. Muscle stimulation was delivered using an implantable miniaturized stimulating device, chosen because of its fixed electrode geometry and ease of insertion into a particular subregion of muscle. Each miniature stimulator was inserted directly into the muscle mass through a large gauge hypodermic needle. Once in the muscle it was used to stimulate motor nerve branches with digitally controlled constant-current stimuli [5], [6]. The biceps femoris (BF) muscle was chosen for the present work because it is a large muscle whose motor-unit organization has previously been described in some detail [12], [13]. The previous studies showed that the muscle has three neuromuscular compart- ments, the posterior, middle, and anterior compartments, which are innervated by separate nerve branches and have different 0018-9294/98$10.00 0 1998 IEEECAMERON et al.: EFFECTS OF REGIONAL STIMULATION USING STIMULATOR IMPLANTED IN POSTERIOR BICEPS FEMORIS 1037 j I sterior Proximal Distal Fig. 1. Line drawing of the experimental preparation showing the location of the distal and proximal devices with respect to anatomical features of BFp. The neural branching pattern is adopted from Chanaud et a[. (1991). The proximal