Is spike frequency adaptation an artefact? Insight from human studies

INTRODUCTION Spike frequency adaptation (SFA) is defined as the decline in motoneuron (MN) firing rate during constant current injection. In classical study of Granit et al. (1963) this phenomenon was investigated in anesthetized animals through injecting MNs with long-lasting rectangular current steps. At least two phases of SFA have been observed in these experiments: the rapid initial phase and the slow (S) late phase (Kernell, 1965). Historically, the first mechanism proposed to explain SFA was the summation of the medium afterhyperpolarization (Kernell, 1972; Kernell and Sjoholm, 1973; Baldissera et al., 1978). However, it has been shown that this mechanism may be responsible only for few initial MN interspike intervals (Powers et al., 1999). More recent studies of the possible mechanisms underlying late SFA have been usually conducted in vitro and often supported by computer simulations (e.g., Sawczuk et al., 1997; Zeng et al., 2005). In these experiments, a multitude of ion channels were blocked with appropriate pharmacological agents and the effects of blocking on SFA magnitude and/or time course were studied. With this type of protocol, certain mechanisms contributing to SFA were identified in non-MN cells. However, in MNs SFA appears to be such a robust phenomenon that blocking of any presumed mechanisms has no effect on its magnitude or time course. Thus, it was concluded that several membrane channels involved in generating rhythmic MN activity contribute to SFA. These mechanisms act together to ensure SFA stability: blocking one set of channels results in an increase in the contribution of the others (Powers et al., 1999; Zeng et al., 2005). One piece of evidence supporting this theory of redundancy was presented by Goh et al. (1989) through experiments involving bullfrog sympathetic neurons. Under normal conditions, selective blockade of the delayed rectifier potassium current (IK) has no effect on SFA. However, SFA would be enhanced through blocking IK, when other IKs (including the calciumdependent current IAHP) were blocked. Thus, the contribution of IK to MN firing patterns depends on the activity of other K+ currents. Considerable part of the research on SFA was produced by Brownstone’s lab (Miles et al., 2005; Brownstone, 2006). Recently, they reported the reversal of SFA during fictive locomotion of decerebrate cat (Brownstone et al., 2011). During the recent meeting of International Motoneurone Community in Sydney the title question of this paper was posed (Brownstone, 2012). The author explains further his concern, asking: “. . . is repetitive firing produced by current injected through the micropipette . . . informative about membrane currents during behaviour? Perhaps SFA is not present . . . during most motor behaviours?” These questions reflect the author’s conviction that a role for late SFA has not been established yet. Human MN studies offer a possibility to investigate intact MNs in their physiological environment. In our opinion, these studies provided enough evidence to answer the questions cited above. This evidence will be presented below. INITIAL SFA The initial SFA phase is present only when the MN starts firing in response to the intracellular rectangular current injection. Since such a rapid depolarization seems to be absolutely non-physiological, this phase is most likely a candidate to be an artefact. However, there is an analogy of rectangular current step in human experiments, when the experimental subject is asked to contract the muscle as quickly as possible (so-called ballistic contractions). This task evokes a sudden increase in synaptic inflow to MN pool, resulting in recruitment of a few MNs with initial firing rates of 60–120 imp/s, which quickly decrease thereafter. The time course of this decrease is similar to those observed in animal MNs in response to constant current injection (Figure 1A). The reaction of the MN to the rectangular current step was investigated by Ito and Oshima (1965). They showed that cat MNs respond to it with an abrupt change in membrane potential, which initially overshoots the target potential, peaks at approximately 15 ms after current onset and stabilizes around 100 ms. The magnitude of the overshoot increased with increasing current intensity. The time course of the initial SFA phase typically observed with rectangular current stimulation appears to follow changes in membrane voltage: the initial firing rate positively correlates with current step amplitude, and the steady level is reached at or shortly after 100 ms (Granit et al., 1963). In contrast, during ramp contractions MN firing rate is gradually increasing until the contraction force reaches