A Neural Pattern Generator that Exhibits Arousal-Dependant Human Gait Transitions

Center for Adaptive Systems and Department of Cognitive and Neural Systems Boston University, 111 Cummington Street, Boston, MA 02215 A neural pattern generator based upon a non-linear cooperative-competitive feedback neural network is presented. It can generate the two standard human gaits: the walk and the run. A scalar arousal or GO signal causes a bifurcation from one gait to the next. Although these two gaits are qua.litatively different, they both have the same limb order and may exhibit oscillation frequencies that overlap. The model simulates the walk and the run via qualitatively different waveform shapes. The fraction of cycle that activity is above threshold distinguishes the two gaits, much as the duty cycles of the feet are longer in the walk than in the run. 1 A Neural Pattern Generator for Human Gait Timing In various quadruped gaits, different relative order of limb oscillations distinguish between various gaits. However, the human walk and run gaits cannot be distinguished on the basis of limb order, since both gaits have the same relative limb order. Nor can they be distinguished on the basis of frequency of oscillation, since each gait may exhibit the same frequency: The limbs may oscillate at the same frequency during fast walk as they do in a slow run. The neural pattern generator described below is an Elli<ts-Grossberg oscillator [2]. It exhibits two distinct oscillatory regimes that can be qmmtit<ttively distinguished on the basis of qualitatively different waveform shapes. Let where xi= -Ax;+ (Bx;)[f(x;) +I;](C + x;)[LD;jg(yj)] y; = E[(ly;)[x;]+yi], [w]+ = max(w, 0) Jil([w]+)2 G1([w]+)2 f(w) = 1~ + ([w]+)2' g(w) = G2 + ([w]+)2' (1) (2)