Introduction

In motor control, firing is the fundamental behaviour of spinal motoneurones being the basis for information transmission between motoneurones and muscle fibers that they innervate. The firing behaviour of motoneurones is controlled by a complex interplay between their intrinsic properties (including plateau potentials) and numerous excitatory and inhibitory synaptic inputs (afferents, inter- and supra-spinal) they receive. One of the important afferent excitatory inputs is the monosynaptic Ia afferent input eliciting the well-known H-reflex, which is widely employed as an instrument for testing of the motoneurone pool excitability (see Schieppati 1987; Knikou 2008; McNeil et al. 2013 for reviews) that is a key intrinsic property in information processing (Rekling et al. 2000; Kernell 2006; Heckman et al. 2009).

In most studies, the motoneurone excitability has been thoroughly investigated in the quiescent state, for non-firing motoneurone pools. However, in normal, motor control is dynamic and repetitive firing is fundamental to the motoneuronal behaviour (Heckman and Enoka 2012). At the same time, as was emphasized by Matthews (1999a), concept of ‘excitability’ should be treated “… with extreme care when it is extrapolated from quiescent neurones to those that are already firing”. In fact, the problem of the estimation of both synaptic excitatory and inhibitory effects on firing motoneurones requires special consideration since motoneuronal information properties, in particular, the somatic membrane potential and the threshold potential for spike initiation essentially change throughout a target interspike interval (ISI) (Calvin 1974; Powers and Binder 1996, 1999). These changes are especially marked within a low discharge rate that is characteristic for the human motoneurone firing behaviour. This low rate is outside of the primary and secondary firing ranges of cat and rat motoneurones thoroughly explored in classical studies by Granit, Kernell, and colleagues (Kernell 1965, 2006; Granit et al. 1996a, b). The peculiarities of this firing range were showed during exploring after-potentials, the excitability and inhibitability of human single firing motoneurones during voluntary muscle contractions and the range has been termed the “sub-primary range” (Kudina and Alexeeva 1992; Kudina 1999). It was found that, during maintaining a postural, the after-hyperpolarization of motoneurones firing in the sub-primary range, in contrast to the primary one, can hardly be regarded as the main factor controlling the human motoneurone firing behaviour. Additionally, both excitability and inhibitability of motoneurones firing within such low firing rate were significantly higher than within the primary range. The possible contribution of the sub-primary range in motor control was further considered in different conditions, e.g., in human motoneurones during studying of force gradation and plateau potential activation (Kiehn and Eken 1997; Matthews 1999b; Wienecke et al. 2009; reviewed by Heckman and Lee 1999; Kernell 2006; Powers and Turker 2010; Duchateau and Enoka 2011) as well as in cat, mouse, and rat motoneurones (e.g., Hamm et al. 2010; Jensen et al. 2018).

Sub-primary range peculiarities can be an important factor in information processing in both normal and diseased human motoneurone pools and, therefore, understanding the issue is urgent. However, the low firing range is difficult to explore, since it is associated with a natural significant variability in ISI duration and influence of the synaptic noise on spike generation and, therefore, is often consciously avoided in animal studies. However, as was emphasized by Calvin and Stevens (1968), noise in the central nervous system (CNS) is the unavoidable consequence of integration of unsynchronized information arriving from various inputs. It is also important to bear being in mind that namely motoneurones firing within the sub-primary range are recruited during natural motor control in humans, in particular, at weak and moderate voluntary muscle contractions, postural tasks, and reflex activations.

In human motor control, motoneurone firing can be explored in during analyzing single motor unit (MU) firing. This method was widely used in many studies (reviewed by Binder et al. 1996; Powers and Binder 1999, 2001; Heckman et al. 2009; Powers and Turker 2010; Heckman and Enoka 2012), including exploring possible evidences of human motoneurone plateau potentials during voluntary muscle contractions (e.g., Kiehn and Eken 1997; Fuglevand et al. 2006; Kudina and Andreeva 2010, 2019; Foley and Kalmer 2019; reviewed by Binder et al. 2020). It was found that the excitability of single firing human motoneurones tested by muscle afferent stimulation (e.g., Buller et al. 1980; Ashby and Zilm 1982; Kudina 1988, 1999; Jones and Bawa 1999) or by cortical volley during transcranial magnetic stimulation (e.g., Boniface et al. 1991; Bawa and Lemon 1993; Olivier et al. 1995) changed, depending first on timing of excitatory volley within a target ISI. The volleys being ineffective in the beginning of the target ISI, became high effective in its end.

This dependence appeared to be characteristic not only in human spinal motoneurones. It was revealed in neocortical neurones studied in slices of the cat sensorimotor cortex (Reyes and Fetz 1993a, b). In these studies, a new interesting aspect of neuronal information transmission was revealed. The authors have firstly shown that in the case of brief transient excitatory volley (duration of 0.5–2.0 ms) its timing within a target ISI largely determined neurone firing probability, resulting in either direct or delayed ISI shortening. A “slow regenerative process” was suggested by the authors as the mechanism underlying the delayed shortening. Note that these interesting data may be considered more widely than only two modes of ISI shortening. They provide important evidence of the possibility of two modes of neurone spiking: usual (direct) and unusual (delayed).

Delayed discharges were also revealed during exploring of rat hippocampus pyramidal neurones in slice preparations (Fricker and Miles 2000) during stimulation evoking small EPSPs. The probability of delayed discharges was reduced as EPSP amplitude was increased. Plateau potential that prolonged the effect of the brief excitatory volley was suggested by the authors as the mechanism underlying the delayed ISI shortening.

It is surprising, but in spite of neurones investigated in reports above, there are still no studies directly showing delayed spiking in animal motoneurones, the most explored neurones in the CNS. To our knowledge, the only study, made on single human MUs (Mattei and Schmied 2002) has previously addressed this issue. However, in this report, both the investigation method (in particular, using the stretch reflex that is no transient testing volley) and result interpretations are rather questionable. As it has been noted by Powers and Turker (2010), data reported by Mattei and Schmied (2002) can reflect different effects evoked by a long EPSP during tendon jerk.

In summary, the question is: the motoneurone delayed spiking is not convincingly revealed until today or it is not characteristic for motoneurones? Thus, as to motoneurones, this question requires further investigations.

The purpose of the present study was looking for possible evidence of delayed spiking evoked by transient excitatory Ia afferent volley in single firing motoneurones of healthy humans.

Methods

The data were collected from four human volunteers (females, aged 47–62 years) who did not have any known neurological disorder. All subjects gave informed written consent for the experimental procedures before the experiments, which was approved by the local Ethics Committee and conformed to the standards of Declaration of Helsinki Experimental protocol.

Experimental protocol

In total, ten experiments were carried out on three muscles: the flexor carpi ulnaris (FCU), the tibialis anterior (TA), and the abductor pollicis brevis (APB). During the experiments, the subjects were seated comfortably in an armchair. They were asked to recruit a few MUs by a weak voluntary isometric contraction of the muscle under study (2–5% of the maximal force) and to maintain MU repetitive firing throughout the experiment, using both visual and auditory feedback of the MU discharges. Individual experiments lasted typically 1–1.5 h; recordings were ceased earlier if the subject found it difficult to continue to control MU firing.

The potentials of single MUs were recorded by selected bipolar needle electrode (leading-off surfaces of 0.015 mm2), inserted in the muscle analyzed in a given experiment, amplified by an electromyograph DISA (Denmark, Type 14 A30) at 200–500 μV/cm, with filters set at 20 Hz–10 kHz and stored on the magnetic tape for off-line analysis. While the subject maintained the muscle contraction, single electrical stimuli of 0.5–1.0 ms duration and interval between stimuli of 1–3 s were applied through bipolar surface electrode to the ulnar nerve (at the elbow) during FCU studies, to the common peroneal nerve (at the head of the fibula) during TA studies, and to the median nerve (at the wrist) during APB studies. In all experiments, stimuli were delivered randomly with respect to MU background repetitive firing. The stimulus intensities were commonly in range of 5–30% from those evoking the maximal H-reflex. Multiple testing was used in all experiments. The main criterion of the choice of the stimulation parameters and number of testing volleys was the full tolerance of the subject to the electrical nerve stimulation in each given experiment to avoid any discomfort for the subject.

Data analysis

The MU action potentials were transferred to a computer by an A/D converter (Russia, type L-Card 154), with a sampling rate of 10 kHz. Action potentials of each single MU were identified on the basis of the waveform shape and amplitude. The results of the preliminary computer identification were then verified by visual inspection of single sweeps by the experienced operator. The trials, in which the MU potentials could not discriminated with certainly, were discarded.

For common statistical estimation of Ia afferent volley effects, peri-stimulus time histograms (PSTHs) of single MUs displaying firing probability before and after the stimulation time (time zero) were constructed. The 200 ms pre-stimulus period of the PSTH was used to estimate the mean number of potentials in a bin (discharge probability) for each MU during its background firing. An increase in discharge probability in a bin of a post-stimulus period of the PSTH was considered to be statistically significant if the number of potentials (counts) in it exceeded a mean number of MU potentials in a background bin by more than two standard deviations (SDs) (P < 0.05).

To estimate the strength of Ia afferent excitatory volley to each MU and MU responsiveness to given excitatory volley, we calculated the common firing index (FI) showing the percentage of MU responses at the H-reflex latency (taken from the PSTH of a given MU as post-stimulus time with significant increasing in firing probability) to the total number of tests, in which a given MU was analyzed.

For estimating Ia afferent volley effects on ISI duration of a firing motoneurone, for each MU, a control background ISI (immediately preceding an each stimulus) and a target ISI (in which testing volley arrived) were calculated for each trial and their distributions for all trials were plotted. The mean values and SDs for control and target ISIs of each MU were calculated and compared, using a paired Student’s t test. Group data are presented in the text as the mean values and their SDs. All differences were typically considered as significant at P < 0.05.

To understand the mechanisms controlling occurrence of the delayed spiking within a target ISI, the previously developed method of testing motoneurone excitability during electrical Ia afferent stimulation evoking the H-reflex of single MUs was used (Kudina 1988, 1999).

In short, since in the experiments, random mode of stimulation in relation to MU background firing and sufficient number of stimuli were used, the afferent volleys appeared rather uniformly throughout the whole target ISI. It made possible analyzing the time course of motoneurone excitability changes within a target ISI by plotting the “interspike-excitability trajectory” for single motoneurone. For this purpose, for each MU, all trials were selected into groups depending on the arrival time of the afferent volley within the target ISI with step of 5–10 ms. For each step, the current FI (showing the percentage of MU evoked responses to the total number of testing volleys arriving in this step of a target ISI) was calculated.

During stimulation, Ia afferent volley always arrives in a target ISI with a delay concerning to stimulation time equal to the H-reflex latency of the MU tested. To determine the timing of the test volley within the target ISI, the time between the last regular voluntary discharge of an MU prior to the stimulus and stimulation point was measured in each trial and the correction was made for the H-reflex latency (taken from the PSTH of the given MU). It should be noted that, because of the MU response latency, just a beginning of a target ISI must be tested by stimuli delivering at the end of a control ISI immediately preceding a target one.

Plotting current FI values versus testing volley timing within a target ISI gave the recovery curve of the motoneuronal excitability after producing a regular voluntary discharge (the interspike-excitability trajectory) characterizing the ability of a motoneurone to fire a spike (direct or delayed) throughout the target ISI.

Results

The essential results are based on recordings from 42 MUs, including 18 MUs in FCU and 24 MUs in TA. Additional results were obtained on five MUs in APB. They were essentially similar to those from FCU and TA.

Common characteristics of Ia afferent stimulation effects on single firing MUs

In all muscles investigated, during weak voluntary muscle contractions, the mean background firing rates of single MUs (estimated as the inverse of mean background ISI values calculated for 0.5–1.0 s before a stimulation time) were close and ranged within the sub-primary range, commonly from 5–6 to 13–15 imp/s.

The typical examples of single MU recordings are presented in Figs. 1 and 2. In number of trials (e.g., Figs. 1a, b and 2a), the stimulation of Ia afferents elicited discharges of firing MUs at the H-reflex latency (characteristic for each muscle), often with some shortening of a target ISI as compared with the control background ISI. If an MU did not fire the short-latency H-response in a given trial, it could fire a discharge at a longer latency, in some trials with a shortened target ISI as well. The instances are presented in Figs. 1b, c and 2b, c. We will provisionally call this response to stimulation at a longer latency the ‘delayed response’ (D-response). For the same MU, D-response latencies were not only longer than those of the H-reflex but also rather variable (e.g., Fig. 2b, c).

Fig. 1
figure 1

Effects of threshold Ia afferent electrical stimulation on single firing MUs of the flexor carpi ulnaris activated by weak voluntary contractions. Asterisks, the stimulation time. H, the MU response at the H-reflex latency; D, the MU delayed response at a longer latency. a Both MUs fired the H-reflex; b One MU fired the H-reflex while the other MU fired the D-response; c Both MUs fired the D-responses. Note, the target ISIs during the D-response were shortened. d H- and D-responses are absent, but both MUs displayed unexpected increase in the target ISI duration. Time bar, 50 ms

Full size image
Fig. 2
figure 2

Effects of threshold Ia afferent electrical stimulation on single firing MUs of the tibialis anterior activated by weak voluntary contractions. Asterisks, H, and D, the same as in Fig. 1. a The MU fired the H-reflex; b The MU fired the D-response; c The non-firing in the background MU was activated at the H-reflex latency, the firing MU fired the D-response. Note, in different trials, the same MU could fire the D-response with changeable latency, but in both cases with shortening target ISIs (b and c). Time bar, 50 ms

Full size image

As can be seen in Figs. 1 and 2, MUs, which fired the H- or D-responses with some shortening of a target ISI, displayed increasing of a post-target ISI as compared with a control ISI. It has been widely accepted that such ISI increasing is explicated by the after-hyperpolarization summation after the evoked response in a firing motoneurone (Calvin and Schwindt 1972). However, in the present experiments, during the motoneurone stimulation by the excitatory volley, some MUs which in a given trial did not fire the H- or D-response could display unexpected increase in the target ISI duration (Fig. 1d).

During stimulation, in addition to responses of firing MUs, sometimes, the excitatory volley could result in the recruitment of a new MU (non-firing in the background) at the H-reflex latency (Fig. 2c). D-response of non-firing MUs was never found. Also, it was never found in isolation, during a very weak stimulation that failed to elicit the H-reflex of firing MUs.

Peri-stimulus time histograms (PSTHs) of single MUs typically revealed the H-reflex and seldom the D-response during Ia afferent stimulation

PSTHs of most MUs (30/42) have revealed the significant increase in firing probability, i.e., the excitatory effect, at the H-reflex latency. The onset latencies for peaks in PSTHs were 18–24 ms for MUs in FCU, 30–38 ms for MUs in TA, and 28–32 ms for MUs in APB. Twelve MUs did not display a significant change in firing probability of PSTHs at the H-reflex latency. These MUs were excluded from further processing.

As to the D-response, it could be revealed as a small but significant increase in discharge probability of one-two bins in a PSTH of some MUs only. The typical instances are presented in Fig. 3. As can be seen, an increase in MU discharge probability in two bins (Fig. 3a) or in one bin (Fig. 3b) of a post-stimulus period of the PSTH were statistically significant (P < 0.05). However, many MUs did not shown the significant excitatory effect at proposed D-response latency (for example, Fig. 3c, d), because of the D-response of the same MU, as was mentioned above, in different trials, displayed changeable latencies and therefore appeared in a few PSTH bins. Moreover, in many PSTHs, D-response timing coincided with some decreasing in firing probability observed after the H-reflex. As a result, the D-response often seemed like a short-living return toward the background level (after the H-reflex) without significant increase in firing probability (Fig. 3c, d). In these cases, we recognized a certain relativity of our assessment of the D-response; however, the further data analysis (in particular, analysis of the target ISI duration) showed that it can be justified. Thus, a PSTH, being commonly the classical method of excitatory effect revealing in a firing neurone, in case of the D-response of human motoneurones, in contrast with the H-reflex, in more cases was no a relevant method.

Fig. 3
figure 3

Peri-stimulus time histograms of single firing MUs of the flexor carpi ulnaris (FCU) and the tibialis anterior (TA) activated by weak voluntary contractions. Abscissa, the time before and after Ia afferent electrical stimulation (zero, stimulation time; the bin width, 3 ms); ordinate, the number of MU potentials (counts) in a bin. Value of mean pre-stimulus count ± 2 SDs are shown by horizontal lines. a, b The significant H-reflex and D-response of the FCU MU (MU mean firing rate, 10.0 imp/s; 491 trials) and the TA MU (7.4 imp/s; 383 trials); c, d The significant H-reflex of the FCU MU (8.7 imp/s; 504 trials) and the TA MU (6.9 imp/s; 485 trials). Significant D-responses are absent

Full size image

Common firing index (FI) as quantitative estimation of MU responsiveness to Ia afferent excitatory volley

As we used the weak Ia afferent stimulation, the common FIs of most single MUs (21/30) were rather small, typically from 6.1 to 25.6% (mean 14.2 ± 5.5). And only nine MUs displayed common FI equal to 30.2–50.8% (mean 40.0 ± 9.0). Depending on FI values, the following tendency was revealed: the marked D-response was found in MUs of the first group only. Thus, in given experimental conditions, the strength of the Ia afferent excitatory volley arriving to single motoneurone influenced on the possibility of the D-response appearance.

Analysis of target ISIs in most MUs revealed significant shortening at evoking both the H-reflex and D-response

Assuming that excitatory volleys arriving in a firing motoneurone should result in some shortening of the target ISI as compared with the control background ISI, we next estimated the duration of target ISIs during evoking the H-reflex and D-response as a possible evidence for or against the two modes of spiking in human motoneurones. For each MU, background control ISIs were calculated. Next, all trials, in which a given MU displayed responses with the H-reflex latency or the D-responses, were selected and mean values of their target ISIs were compared with mean values of control ISIs of a given MU. It has been found that although in some trials target ISIs displayed no shortening during firing the H-reflex and D-responses, but for most MUs (15/17) mean target ISI was significantly decreased, compared to the mean control ISI for both evoked responses (P < 0.05). The typical instances of ISI distributions are presented in Fig. 4. Consequently, the target ISI analysis confirmed the validity of our assessment of the D-responses and showed that the present of some delayed responses in the PSTH can be ‘concealed’ by their various latencies.

Fig. 4
figure 4

Distributions of control and target interspike intervals (ISIs) of the single firing MU of the flexor carpi ulnaris (FCU) and that of the tibialis anterior (TA) during Ia afferent electrical stimulation (493 and 750 trials, respectively). Abscissa, the duration of control and target ISIs (ms), the bin width, 10 ms; ordinate, number of ISIs in each bin (as percent of common number of ISIs). a, b Gray bars, control ISIs (102.9 ± 20.5 ms for the FCU MU and 138.7 ± 19.6 ms for the TA MU), black bars, the H-reflex target ISIs (97.2 ± 18.7 ms for the FCU MU and 127.9 ± 23.3 ms for the TA MU); c, d Gray bars, the same as in a, b; black bars, the D-response target ISIs (95.6 ± 18.7 ms for the FCU MU and 124.3 ± 16.8 ms for the TA MU). Differences between the duration of the control and target ISIs during evoking the H-reflex and D-response for both muscles were statistically significant (P < 0.05)

Full size image

Effects of Ia afferent excitatory volley on target ISI duration of single firing motoneurone depended on volley timing within target ISI

As is known from the current literature (see “Introduction”), that effects of excitatory and inhibitory volleys on firing motoneurone depend from volley arrival moment (timing) within an ISI. To estimate timing of excitatory volley within a target ISI resulting in the occurrence of the H-reflex or D-response, we turned to analyses of changes in the duration of the target ISIs at different timing of the excitatory volley throughout a target ISI.

In our experimental conditions, we observed that the target ISI duration widely depend from timing of the afferent excitatory volley within the target ISI. In most trials, at timing of the afferent volley in the beginning of a target ISI, the latter appeared to be either lengthened or unchanged in compared to the control ISI (Fig. 5). When the excitatory volley arrived in the later part of the target ISIs, some of them demonstrated shortening, as compared to control ISIs, while the other did not. Note that shortening of the target ISIs was found during evoking both the H-reflex (see Fig. 5, the points on diagonal line) and the D-response (Fig. 5, the triangles). In the end of the target ISIs, the afferent volley did not evoke visible changes in the target ISIs, even if the MU fired the response at the H-reflex latency. Thus, during the estimation of the excitatory effect by target ISI shortening alone, it is possible that some H-reflexes and D-responses, which could appear without target ISI shortening, could be underestimated.

Fig. 5
figure 5

Changes in target ISI duration of the single firing MU of the flexor carpi ulnaris (FCU) and that of the tibialis anterior (TA) to threshold Ia afferent volley throughout a target ISI. Abscissa and ordinate, arrival time of testing volley within the target ISI and duration of the target ISI (as percent of control ISI), respectively. Arrows, limits of scatter in background ISIs (mean ± 2 SDs). Each sign is the result of one trial. The open circles, responses at the H-reflex latency; the triangles, the D-responses; filled circles, the remaining target ISIs. The FCU MU and TA MU, 443 and 606 trials, respectively

Full size image

Revealed prolonged ISIs during afferent volley arrival in the earlier part of the target ISIs were discarded from further analysis, because they obviously could not be evoked by the excitatory volley and therefore their analysis is well beyond the scope of this report. Note only that similar unexpected prolonged ISIs during transient excitatory volley arriving in the beginning of the target ISI were recorded in the study of Reyes and Fetz (1993b) and were named by the authors “paradoxical” effect. This interesting phenomenon requires a special research.

Interspike-excitability trajectories allow quantitative estimating of single firing motoneurone excitability to Ia afferent volley throughout target ISI

To estimate quantitatively the excitability changes of each motoneurone throughout the target ISIs, we calculated the current FIs. It was found that Ia afferent volleys, arriving within the first part of the target ISI (typical range, 35–60%; sometimes up to 75%) usually failed to elicit a motoneurone discharge at the H-reflex latency (the current FI was equal 0%). As the timing of the Ia afferent volley within the ISI increased, the number of motoneurone responses at the H-reflex latency also increased and in the end of a target ISI the MU current FI could reach 100%. In contrast, the same MUs started to display the D-response during afferent volley timing in the first half of a target ISI (typical range, 20–55%). Then, within following part of the target ISI, when the motoneurone displayed the maximal FIs during the H-reflex evoking, the possibility of the D-response occurrence decreased and in the end of the target ISI it failed to occur. Typical results are shown in Fig. 6. Thus, the effective timing at which the firing motoneurone was ready to fire a discharge in response to Ia afferent excitatory volley at the H-reflex and D-response latency, essentially appeared to be surprisingly different. The D-response occurred at earlier timing of excitatory volley and its current FI never reached 100%.

Fig. 6
figure 6

Interspike-excitability trajectories as quantity estimation of excitability of single firing MU of the flexor carpi ulnaris (FCU) and that of the tibialis anterior (TA) to threshold Ia afferent volley throughout the target ISI (a and b, respectively). Abscissa, arrival time of testing volley within the target ISI as percent of control ISI (step, 10%); ordinate, current firing index as percent of MU H- or D-responses to the total number of trials arriving in each ISI step. Open circles, responses at the H-reflex latency; triangles, the D-responses. The FCU MU and TA MU, 443 and 750 trials, respectively

Full size image

Discussion

The present results provide some evidence of the appearance of two types of spiking (direct and delayed) evoked by weak electrical stimulation of Ia afferents in spinal motoneurons discharging at low discharge rates, in the sub-primary firing range, during voluntary muscle contractions in healthy humans. The first type of evoked responses was the well-known H-reflex while the second, unusual one, named the D-response, was characterized by essentially different properties. The main differences were the following. The D-response had latencies that were longer and considerably more variable than the H-reflex latencies. The mode of motoneurone spiking was found to be correlated with Ia afferent excitatory volley strength and testing volley timing within a target ISI. The D-response was found during testing by weaker excitatory volley evoking the H-response with lower common FIs. At stronger excitatory volleys, D-response did not appear. The important property was particular timing of excitatory volley within a target ISI resulting to the D-response.

These properties of the D-response of human motoneurones strongly resemble those of delayed discharges reported by Reyes and Fetz (1993a, b) during intracellular recording in cat neocortical neurones. The authors have described two modes of ISI shortening resulted from direct and delayed crossings firing level by transient excitatory volley depending on its strength and timing in the target ISI. It has been shown that spikes evoked by delayed crossing (resulting from EPSPs arriving in the earlier part of the target ISI) had latencies that were longer and more variable than latencies associated with direct crossing (during the EPSP arriving toward the end of the ISI). The stronger direct response, the weaker delayed response appeared to be. It seems unlikely that the similarity of our findings to these results is fortuitous.

In relation to the mechanisms underlying the delayed spiking, Reyes and Fetz (1993a) suggested the activation of a slow regenerative process, which may activate the sodium persistent inward current (PIC) in cat neocortical neurons. Fricker and Miles (2000) proposed as well sodium PIC that underlies the plateau potentials as the mechanism of the delayed spiking in rat hippocampal neurons. It is possible the plateau potentials can be included as a mechanism underlying the delayed spiking in human motoneurones. This mechanism can be assumed, since MU firing behaviour found in our experiments exhibited patterns consistent with the behaviour of neurons in the above studies. Note that since no methods are available for direct estimating these effects in humans, further careful examination of single MU firing patterns might be a valuable approach.

Delayed spiking evoked by a small excitatory volley can be more important than it appears at first. In normal motor control, in contrast with lab experiments, weaker excitatory volleys, resulting in delayed responses, can be more relevant natural mode of neurone activation. According to Calvin (1975), “While experimenters often synchronize many inputs to produce a large compound PSP which can cross threshold, this artificial situation provide little physiological insight into how temporal pattern of inputs produce trains of output spikes.”

Understanding the delayed spiking underlying mechanisms is especially important in the clinic. Some clinical data have led us to assume that, sometimes, motor control disturbances can result in essential delayed spiking. For example, it has been reported (Das Gupta 1963) that in Parkinsonism a prominent feature of the firing pattern of single MUs during voluntary muscle contractions was a tendency to appearance of unusual MU firing (called paired discharges) with extremely variable ISIs, between 20 and 80 ms, but commonly from 30 to 60 ms. The similar MU firing with unusual ISIs ranged between 40 and 50 ms was reported by Anastasopoulos (2020) during exploring tremor mechanisms in Parkinson’s disease. The underlying mechanisms of this unusual MU firing are not quite clear. However, it may be suggested that such motoneurone firing behaviour may include increased delayed spiking as one of the mechanisms.

As was noted in “Introduction”, in the current literature, delayed spiking has not been reported in animal spinal motoneurones. The only work devoted to this topic was the study by Mattei and Schmied (2002). The authors reported so-called “prolonged and delayed spiking” which the authors named also “advanced spiking” of human MUs evoked by tendon jerk. Authors noted that “the stronger the spike-triggering effect, the stronger the advanced spiking effect turned out to be”. However, it should be emphasized that delayed spiking studied by Reyes and Fetz (1993a, b) as well as Fricker and Miles (2000) exhibited quite opposite relation. In addition, excitatory volley, as was reported by Mattei and Schmied (2002), was lasting and shortened not only a target ISI but as well one or even two following ISIs, in contrast with the data by Reyes and Fetz (1993a, b) exploring “brief, transient” volley. It is well known that the brief volley affects only the ISI in which it occurs and after a spike motoneurone “forgets” this volley, starting a new cycle (Calvin and Stevens 1968; Fetz and Gustafsson 1983). Therefore, we concluded, in agreement with Powers and Turker (2010), that data reported by Mattei and Schmied (2002) could reflect quite different effects, obviously evoked by a long EPSP during tendon jerk.

The next important question is what could be the most relevant estimation of the responsiveness of firing motoneurones to an excitatory volley? A responsiveness measure for neocortical neurones, so-called the phase-response curve, was introduced in the reports of Reyes and Fetz (1993a, b), during studying cat neocortical neurones firing and then was used in a number of other studies (e.g., Gutkin et al. 2005; Farries and Wilson 2012; Goldberg et al. 2013; Stiefel and Ermentrout 2016). During plotting the phase–response curve (commonly with using trials with marked shortening of a target ISIs), it had typically zero values at the start and end of the target ISI and a peak in the middle (Reyes and Fetz 1993a, b; Gutkin et al. 2005). However, the similar estimation is hardly acceptable for the analysis of human firing motoneurones during natural motor control at weak isometric voluntary muscle contractions and postural tasks because of MU low firing rates within the sub-primary range (usually 5–15 imp/s) and high discharge variability. As a result, during both mode of motoneurone spiking (direct and delayed), evoked by small transient excitatory volley, ISI shortenings tend to be small and could be occluded by motoneurones’ normal discharge variability. Therefore, in many trials, in response to the excitatory volley, a motoneurone fired the spike with small or even no statistically significant shortening of target ISIs. However, this complicity of estimation of the excitatory volley effect was characteristic for both responses: the direct response (the H-reflex) and the D-response. Consequently, the absence of marked ISI shortening does not mean the absence of excitatory effect, as the well-known H-reflex is the result of Ia afferent excitatory volley (according to the PSTH data). Therefore, at the estimation of excitatory effects arriving in a human motoneurone during natural motor control additional criteria are needed.

We assume that in natural motor control the most functionally significant measure of the motoneurone excitability is the MU discharge occurrence itself, independently from value of target ISI shortening, as motoneurone spiking is a primary key property determining to be or not to be the muscle contraction, while ISI shortenings are secondary. Therefore, we suppose that interspike-excitability trajectory (characterizing the ability of a motoneurone to fire spike throughout the target ISI in response to an excitatory volley) is the most relevant estimation of the excitability of a human firing motoneurone during natural motor control. Earlier, this criterion has been used for the estimation of the excitability and inhibitability of the human firing motoneurones in different conditions (Kudina, 1988, 1999; Kudina and Andreeva, 2017). Adequacy of similar estimation as the more functionally relevant, since namely it determines capability of motoneurone to respond to a given synaptic input, was noted by Matthews (1996). Furthermore, in rat and cat motoneurone investigations, it has been concluded that rate modulation will tend to play a lesser role than motoneurone recruitment for low force gradation, e.g., during postural tasks (Bakels and Kernell 1994; Kernell et al. 1999).

In conclusion, while the present results provide some evidences allowing supposing that delayed spiking exists in human firing motoneurones during natural muscle contractions, however, in this case, the possibilities of the indirect analysis of information processing in human motoneurones gives no definitive proof. Until incontestable intracellular data in animal motoneurones are obtained, the possibility of two modes of spiking in motoneurones should be considered only as hypothetical. Further studies of this interesting issue in animal motoneurones are very desirable. As was underscored by Duchateau and Enoka (2011), progress in understanding of critical issues of motor control “will depend on the successful integration of observations derived from studies on experimental animals and humans”.