Search

In the operant conditioning experiment described by HEFFERLINE et al. (1959) a subject terminates or postpones an aversive noise stimulation by producing a particular myoelectrical activity (operant response : OR). These authors observe an increase of OR occurrence frequency in subjects ignoring the aim of the experiment (naive subjects), and they come to the conclusion that this method is efficient. Nevertheless, the validity of this conclusion is questioned by the lack of control experiments. Hence, this work has to be restudied. We have submitted 13 subjects to a control experiment. Subjects were placed under conditions similar to those used by HEFFERLINE et al., but without any contingency relating the myoelectric activity with the aversive stimulus. The results of these open-loop experiments showed that in a given subject, OR occurrence frequency may vary in large proportions as a function of time. In HEFFERLINE's naive subjects, the increase of occurrence frequency remained completely within the limits of spontaneous variations observed in our control subjects. Therefore, this increase is not significant. We have submitted 18 subjects to a conditioning experiment. 14 subjects among them produced more OR than those of the control subjects during the closed-loop experiment. We conclude that these 14 subjects have been actually conditioned.

Inspiratory duration (TI), cycle duration (TTOT), and tidal volume (VT) were continuously measured in 11 normal subjects during 400 respiratory cycles. Small breath to breath changes in these variables were separately analyzed. For each of these variables, successive observations are not statistically independent; "large" values tend to be followed by "large" values. A respiratory feedback may be involved in this sequential dependence. In that case, any known system of respiratory control could be associated with it, even those with time constant or delay longer than one cycle duration.

If myoelectrical activity is joined with some reinforcement in a biofeedback loop, naive subjects are able to learn a particular myoelectrical response. This learning is established by the increase of operant response occurrence frequency. Data resulting from control experiments are used as a reference.