No announcement yet.


This topic is closed.
  • Filter
  • Time
  • Show
Clear All
new posts

  • FES/FNS and ALNs/NNs

    Dear Biomch-L readers,

    In my posting of 22 August on FES & ALN, I wrote that Prof. William W.
    Armstrong at the University of Alberta was sending me some (p)reprints
    on the application of his Adaptive Logic Networks (ALNs) to Functional
    Electrical Stimulation (FES). Today I received this material and some
    other reprints from him; it is a pleasure to quote in full the abstract
    of the nominated "best student award paper" for the upcoming 14th Annual
    International Conference of the IEEE Engineering in Medicine and Biology
    Society in Paris, 29 Oct - 1 Nov 1992:

    Alexander Kostov, Richard B. Stein, William W. Armstrong~ & Monroe Thomas~
    Division of Neuroscience, and ~Computing Science Department,
    513 HMRC, University of Alberta, Edmonton, AB, Canada, T6G-2S4

    Abstract - An Adaptive Logic Network (ALN), a type of Neural Network (NN),
    was evaluated for the control of walking in Spinal Cord Injured (CSI) pa-
    tients. The motivation behind this research was to explore more reliable
    methods for control of simple Functional Neuromuscular Stimulation (FNS)
    systems in incomplete SCI patients. The ALN was used to recognize a pa-
    tient's intention to make a step by stimulating muscles in a partially
    paralyzed leg. Signals from four force sensors, installed under the toes
    and heels, have been used as inputs and a gating pulse associated with
    the stimulation as an output for learning and testing the function of the
    control system. Manual control, by either the patient or a physiothera-
    pist, has been used as a template to be matched by the ALN. Generaliza-
    tion of the learned functions by the ALN to previously unseen data was
    also tested. Finally, we manipulated the number of input channels, the
    inclusion of information from past samples and prediction of future
    events. The ALN is capable of generating the same time series of output
    pulses as those generated by "human experts". Furthermore, it can predict
    the stimulation event early enough so that the requirement for stimulation
    can be verified and the patient informed to prepare for stimulation.

    Personally, I find the final caution in this paper quite relevant:

    " ... (The ALN) program could be a very valuable tool in signal proces-
    sing, pattern recognition and generating rules for control of gait for SCI
    patients. However, the safety of its performance has not been verified
    under a wide range of conditions and until this has been accomplished the
    ALN must be combined either with algorithmic restrictions or expert sys-
    tems that will prevent its unsafe operation."

    The reference for an already published (and invited) paper is:

    R.B. Stein, A. Kostov, M. B'elanger, W.W. Armstrong and D.B. Popovi'c
    (1992), Methods to control Functional Electrical Stimulation in Walking.
    First International FES Symposium, Sendai, Japan, July 23-25, 1992

    Abstract - We have compared the performance of simple rule bases (sets of
    rules governing transitions between states) and adaptive logic networks
    (a type of neural network) in predicting a desired pattern of electrical
    stimulation. Data were obtained either from force sensors in the shoes of
    a patient walking with functional electrical stimulation or sensory nerves
    supplying the foot of a cat that was walking on a treadmill. The rule base
    performed reliably in setting the times of stimulation in the patient and
    the cat. The neural network could learn, in addition, the entire pattern
    of EMG activity from the signals recorded in the cat.

    Regards -- hjw