Does trans‐spinal and local DC polarization affect presynaptic inhibition and post‐activation depression?

Abstract

Key points

• Trans‐spinal polarization was recently introduced as a means to improve deficient spinal functions. However, only a few attempts have been made to examine the mechanisms underlying DC actions. We have now examined the effects of DC on two spinal modulatory systems, presynaptic inhibition and post‐activation depression, considering whether they might weaken exaggerated spinal reflexes and enhance excessively weakened ones.

• Direct current effects were evoked by using local intraspinal DC application (0.3–0.4 μA) in deeply anaesthetized rats and were compared with the effects of trans‐spinal polarization (0.8–1.0 mA).

• Effects of local intraspinal DC were found to be polarity dependent, as locally applied cathodal polarization enhanced presynaptic inhibition and post‐activation depression, whereas anodal polarization weakened them. In contrast, both cathodal and anodal trans‐spinal polarization facilitated them.

• The results suggest some common DC‐sensitive mechanisms of presynaptic inhibition and post‐activation depression, because both were facilitated or depressed by DC in parallel.

Abstract

Direct current (DC) polarization has been demonstrated to alleviate the effects of various deficits in the operation of the central nervous system. However, the effects of trans‐spinal DC stimulation (tsDCS) have been investigated less extensively than the effects of transcranial DC stimulation, and their cellular mechanisms have not been elucidated. The main objectives of this study were, therefore, to extend our previous analysis of DC effects on the excitability of primary afferents and synaptic transmission by examining the effects of DC on two spinal modulatory feedback systems, presynaptic inhibition and post‐activation depression, in an anaesthetized rat preparation. Other objectives were to compare the effects of locally and trans‐spinally applied DC (locDC and tsDCS). Local polarization at the sites of terminal branching of afferent fibres was found to induce polarity‐dependent actions on presynaptic inhibition and post‐activation depression, as cathodal locDC enhanced them and anodal locDC depressed them. In contrast, tsDCS modulated presynaptic inhibition and post‐activation depression in a polarity‐independent fashion because both cathodal and anodal tsDCS facilitated them. The results show that the local presynaptic actions of DC might counteract both excessively strong and excessively weak monosynaptic actions of group Ia and cutaneous afferents. However, they indicate that trans‐spinally applied DC might counteract the exaggerated spinal reflexes but have an adverse effect on pathologically weakened spinal activity by additional presynaptic weakening. The results are also relevant for the analysis of the basic properties of presynaptic inhibition and post‐activation depression because they indicate that some common DC‐sensitive mechanisms contribute to them.

Key points

• Trans‐spinal polarization was recently introduced as a means to improve deficient spinal functions. However, only a few attempts have been made to examine the mechanisms underlying DC actions. We have now examined the effects of DC on two spinal modulatory systems, presynaptic inhibition and post‐activation depression, considering whether they might weaken exaggerated spinal reflexes and enhance excessively weakened ones.

• Direct current effects were evoked by using local intraspinal DC application (0.3–0.4 μA) in deeply anaesthetized rats and were compared with the effects of trans‐spinal polarization (0.8–1.0 mA).

• Effects of local intraspinal DC were found to be polarity dependent, as locally applied cathodal polarization enhanced presynaptic inhibition and post‐activation depression, whereas anodal polarization weakened them. In contrast, both cathodal and anodal trans‐spinal polarization facilitated them.

• The results suggest some common DC‐sensitive mechanisms of presynaptic inhibition and post‐activation depression, because both were facilitated or depressed by DC in parallel.

Abbreviations

DH

dorsal horn

locDC

local direct current

MN

motor nucleus

PAD

primary afferent depolarization

Per

peroneal

Q

quadriceps

Sur

sural

T

threshold

tsDCS

trans‐spinal direct current stimulation (or trans‐cutaneous spinal direct current stimulation according to another convention)

Introduction

Presynaptic inhibition and post‐activation depression provide the two main means of control of input to spinal motoneurons, and various motor disorders might result from their deficiencies. For instance, exaggerated stretch reflexes in spastic patients were found to be associated with abnormally weak presynaptic inhibition as well post‐activation depression of synaptic actions of group Ia afferents (Pierrot‐Deseilligny & Burke, 2005; Nielsen et al. 2007; Grey et al. 2008; Andrews et al. 2015; Hedegaard et al. 2015). Abnormally strong presynaptic inhibition and post‐activation depression might, in contrast, contribute to the weakening of synaptic input to spinal motoneurons found in recent studies of spinal muscular atrophy by Mentis et al. (2011). Modulation of the degree of presynaptic inhibition and post‐activation depression might thus be beneficial in some of these cases, and the potential of non‐invasive trans‐spinal DC application to this end has already been considered. However, the possibilities of beneficial effects of DC on spinal cord activity have so far been demonstrated in only very few cases, e.g. in studies of human H reflexes (Winkler et al. 2010; Knikou et al. 2015), of pain threshold (Cogiamanian et al. 2012; Knikou et al. 2015) and responses to stimulation of skin afferents after spinal cord injury (Hubli et al. 2013), and primarily in healthy subjects (see Priori et al. 2014). In addition, when the effects of trans‐spinal direct current stimulation (tsDCS) on post‐activation depression were estimated in humans, considerable interindividual variations were found, even though the dominant effect was the enhancement of H reflexes by cathodal tsDCS and their weakening by anodal tsDCS (Winkler et al. 2010; Lamy et al. 2012; Knikou et al. 2015).

In the present study, the effects of DC on presynaptic inhibition and post‐activation depression were examined in acute experiments in animals. The animal preparation was chosen because it allows DC to be applied locally (locDC) and thereby permits the analysis of much more spatially restricted DC effects than those evoked during trans‐spinal polarization. As demonstrated previously, the intensity of the effective locDC may be reduced to <0.2–0.3 μA as compared with the intensity of tsDCS, which is of the order of milliamperes (Bolzoni & Jankowska, 2015; Jankowska et al. 2016).

The rationale for examining DC effects on presynaptic inhibition was that presynaptic inhibition is associated with a potent depolarization of afferent fibres (primary afferent depolarization; PAD), which might be either enhanced or counteracted by DC. Local cathodal polarization and the resulting depolarization of the fibres would be expected to decrease the amount of transmitter released from these afferents (Eccles et al. 1962 a) and, therefore, to enhance the PAD‐related presynaptic inhibition, whereas anodal polarization would have the opposite effects (for a review of effects of PAD, see Rudomin & Schmidt, 1999; and for a review of effects of other factors, see Pelletier & Cicchetti, 2015). However, it was also an open question whether DC would induce as long‐lasting changes in the presynaptic inhibition as in the excitability of the afferents (Bolzoni & Jankowska, 2015; Jankowska et al. 2016).

The rationale for examining DC effects on post‐activation depression was similar. Post‐activation depression was found in human subjects as well as in animals (for references, see Crone & Nielsen, 1989; Hultborn et al. 1996), and it was demonstrated that the release of transmitter from Ia terminals following their activation is reduced for several seconds (Hultborn et al. 1996). Such a reduction might thus be enhanced by long‐lasting depolarization of presynaptic terminals, which would by itself reduce the probability of transmitter release (Hagiwara & Tasaki, 1958; Eccles et al. 1962 a; Hubbard & Willis, 1962 a,1962 b).

The main aim of this study was therefore to examine the long‐lasting postpolarization effects of locally applied DC on presynaptic inhibition and on post‐activation depression. The second aim was to compare the effects of locDC and tsDCS and to determine which effects of locally applied cathodal and anodal DC in experiments on animals might underlie the effects of tsDCS in humans.

The generally similar effects of DC on presynaptic inhibition and on post‐activation depression found during the first stages of this study have provided new indications of their mechanisms. The third aim of the study became, therefore, to analyse to what extent presynaptic inhibition and post‐activation depression might depend on the same DC‐sensitive mechanisms.

Methods

Ethical approval

All experiments were approved by the Regional Ethics Committee for Animal Research (Göteborgs Djurförsöksetiska Nämnd) and followed National Institutes of Health and EU guidelines for animal care and the ethical principles of The Journal of Physiology (Drummond, 2009; Grundy, 2015), complying with the Journal's ethics list. The animals were bred and housed under veterinary supervision at the Laboratory of Experimental Biomedicine at Sahlgrenska Academy, where the experiments were carried out.

Preparation

The experiments were performed on 67 adult rats of both sexes (Wistar, 350–600 g). Anaesthesia was induced with isoflurane (Baxter Medical AB, Kista, Sweden; 4% in air), followed by administration of α‐chloralose (Rhône‐Poulenc Santé, France; 60 mg kg⁻¹, i.p.) and pentobarbital sodium (Apoteksbolaget, Göteborg, Sweden; 15 mg kg⁻¹ i.p.). During the course of the experiments, the anaesthesia was supplemented with three or four additional 20 mg kg⁻¹ doses of α‐chloralose up to 120–160 mg kg⁻¹, at 1–2 h intervals, as well as when the continuously monitored heart rate increased above 400 beats min⁻¹. The preliminary dissection included tracheal intubation, a tail vein cannulation, insertion of an i.p. catheter, dissection of the sural and peroneal nerves, and sometimes also of the femoral nerve, as well as exposure of the L2–L5 spinal segments by laminectomy. Paraffin oil pools were constructed by skin flaps above the dissected tissues. During later stages of the experiments, the neuromuscular transmission was blocked by pancuronium bromide (Pavulon; Jelfa, Jelenia Gora, Poland) applied i.v. (via the tail vein), in an initial dose of 0.3 mg kg⁻¹, supplemented when needed with 0.1 mg kg⁻¹. Pancuronium bromide was applied ∼3 h after the induction of the anaesthesia when its depth was verified to be stable. Artificial respiration was then applied (80–90 breaths min⁻¹, 0.3–0.6 ml min⁻¹) using a respiratory pump (model SAR‐830/P; CWE, Ardmore, PA, USA) to maintain the expired CO₂ level at 3–4%.

The core body temperature was maintained at ∼38°C by servo‐controlled heating lamps. In order to compensate for fluid loss and to prevent the deterioration of the state of the animals, 10–20 ml of acetate buffer was injected s.c. and 0.5–0.7 ml of the plasma expander Dextran 40 (Rheomacrodex, Meda AB, Solna Sweden) was administered i.v. at the beginning of the experiments. The experiments were terminated by a lethal dose of pentobarbital followed by excision of the heart.

Recording

The effects of DC application on presynaptic inhibition were estimated from changes in extracellular monosynaptic field potentials evoked from skin afferents in the dorsal horn and from group Ia muscle afferents in the motor nucleus when preceded by conditioning stimuli applied to another nerve. The effects of DC on post‐activation depression were examined by comparing monosynaptic field potentials evoked in the same way, but preceded by stimulation of the same nerve at time intervals at which only negligible presynaptic inhibition was evoked. For timing and parameters of the stimuli, see next subsections. The field potentials were recorded with glass micropipettes filled with a 2 m solution of NaCl (tip ∼2 μm, impedance 1.5–5 MΩ) and a conventional high‐impedance amplifier (low‐pass filters 15 or 1 Hz, high‐pass filter 5 or 3 kHz). Afferent volleys were recorded with a silver–silver chloride ball electrode in contact with the surface of the spinal cord at the L2 spinal level against a reference electrode inserted into the back muscles at the same segmental level and as a tri‐phasic or di‐phasic potential preceding the extracellularly recorded field potentials. Both the original records and averages of records evoked by 10 stimuli at ∼0.2 Hz were stored online.

Stimulation

The sural and peroneal nerves were stimulated via double silver–silver chloride electrodes in a paraffin oil pool, and the femoral nerve was stimulated by an indwelled tunnel electrode. Constant‐voltage current pulses of 0.2 ms duration were used to this end, with the intensity of the stimuli being expressed in multiples of threshold stimuli (as defined by records of incoming volleys recorded from the surface of the spinal cord). The peroneal nerve was stimulated at 1.5–1.8 times threshold (T) stimuli to activate group I afferents and at 5T to activate skin and group II afferents. The femoral nerve was stimulated at 2–2.5T to activate group I afferents (aiming at group Ib afferents in the quadriceps nerve). The sural nerve was stimulated with 2–3T stimuli in order to activate most of the largest skin afferents in this nerve. When the peroneal nerve was stimulated at 5T, no attempts were made to differentiate between effects evoked from skin and grou  II afferents in this nerve, because both may contact the same neurons in the dorsal horn (Edgley & Jankowska, 1987), both induce presynaptic inhibition of cutaneous afferents (Eccles et al. 1962 b) and both display post‐activation depression (Hammar et al. 2002).

Pairs of a recording glass micropipette and a stimulating tungsten electrode were mounted in two holders of a double‐headed manipulator operated by separate step motors, allowing the final intraspinal placement of the two electrodes within a preselected distance from each other (Engberg et al. 1972; see Fig. 1 C). While both electrodes were above the spinal cord, their tips were aligned at a distance of ∼5 μm from each other; thereafter, the tungsten electrode was withdrawn by a known distance. This allowed the identification of the optimal recording site using the microelectrode before the tungsten electrode was returned to its original position. The electrodes were inserted into the spinal cord through the dorsal columns, through a ∼1 mm² hole made in the dura mater and with the pia mater peeled away from the surface of the spinal cord between blood vessels. This enabled access to deeper structures with minimal damage caused by pressure. The final position of the microelectrodes within a motor nucleus was guided by recordings of antidromic field potentials evoked by stimulation of a muscle nerve and/or of monosynaptic field potentials from the group I muscle afferents, as indicated by the threshold of the field potentials as low as that of the afferent volleys. These locations (at 0.9–1.20 mm depth from the surface of the spinal cord, at an angle of 2–10 deg, tip directed laterally) were defined using the experimental set‐up in Fig. 1 A. In the dorsal horn, the electrodes were positioned at the site (at ∼0.6–0.7 mm depth from the surface of the spinal cord; Fig. 1 B) where the largest monosynaptic field potentials were evoked by stimulation of low‐threshold (1.1–1.6T) skin afferents in the sural nerve and/or by higher‐threshold (3–5T) afferents in the peroneal nerve. The responses of the postsynaptic neurons activated by these afferents provided an additional indication that the electrode tip was close to the target cells of the afferents.

Figure 1. Experimental design for tests involving local DC application.

A and B, diagram of the set‐up used to examine the effects of DC on postsynaptic potentials evoked by muscle (peroneal) and cutaneous (sural) afferents. The postsynaptic potentials (extracellular field potentials) were recorded with a glass micropipette in the peroneal motor nucleus and in the dorsal horn, respectively, while a tungsten electrode was used to deliver the polarizing current as well as for intraspinal stimulation. C, double‐headed manipulator for placement of the two electrodes, the glass micropipette in the holder A and the tungsten electrode in the holder B. D, the timing of the stimuli used to induce the tested responses with respect to the polarization periods. Arrows indicate the periods during which the sequences of these responses were evoked (three times during the 10 min control period; once during the successive 5 min periods). E and F, the timing of the test and conditioning stimulus sequences within each of these periods. When both post‐activation depression and presynaptic inhibition were tested during the same periods, the sequences of these stimuli were applied one after another. [Color figure can be viewed at wileyonlinelibrary.com]

Local polarization

The polarization was applied using a custom‐designed, battery‐driven, constant‐current stimulator (D. Magnusson, University of Gothenburg). The stimulator supplied a continuously monitored current within a range of intensities of 0–1 μA. The current was passed via a tungsten electrode placed in a double‐headed manipulator, as described in the preceding subsection, against a 2‐cm‐long tungsten electrode inserted into the back muscles in the mid‐line along the vertebral column close to the rostral edge of the laminectomy, or against a crocodile clip attached to the first spinous process rostral to the laminectomy. The polarizing current was applied for a total of 25 min, with five applications of 5 min per application, separated by 5 min intervals, because such polarization timing greatly increases the probability of the development of long‐lasting postpolarization effects (Bolzoni & Jankowska, 2015). Effects of local DC polarization were previously found when the intensity of the current was within the range of 0.1–0.5 μA; however, 0.2–0.3 μA intensity was routinely used in the present study to avoid risking failures due to too weak polarization or cathodal block that was sometimes caused by ≥0.5 μA stimuli.

Trans‐spinal polarization

Trans‐spinal polarization was applied using another constant‐current stimulator, supplying current within a range of intensities of 0–2 mA. The current passed between a crocodile clip attached to the dorsal processes of the L1 or Th13 vertebrae, 8–15 mm rostral to the recording microelectrode, and another crocodile clip attached to the skin flap at the abdomen at the level of the last rib, i.e. approximately at the same rostrocaudal level across the spinal cord. The intensity of the DC was selected based on the effects of various current intensities on the excitability of nerve fibres stimulated within the dorsal horn. Changes in their excitability were estimated from increases or decreases in the areas of earliest components of compound action potentials recorded from a peripheral nerve (see Results, Fig. 5). The minimal effective current was ∼0.2 mA, but the effects increased when the intensity was increased up to 1.1 mA. Accordingly, we selected intensities of 0.8–1.0 mA as effective but still submaximal.

Figure 5. Experimental design for tests involving application of trans‐spinal DC stimulation (tsDCS).

A, diagram of the set‐up used to examine the effects of tsDCS on the excitability of primary afferents activated by intraspinal stimuli. The stimuli were delivered via a tungsten electrode positioned within a region in which large monosynaptic field potentials evoked by these afferents were previously recorded, as depicted in Fig. 1 A–C. The tsDCS was applied between an electrode attached above the vertebral column and a reference electrode in contact with the skin on the abdomen. B, examples of compound action potentials recorded from the peroneal nerve, evoked by 10 μA stimuli before, during and after tsDCS, as indicated above and below the records. Changes in the excitability of the stimulated fibres were estimated from the areas of the early components of the potentials (boxed). C, plots of the areas of these action potentials (ordinate) during cathodal and anodal tsDCS of increasing intensity (abscissa). Note that cathodal tsDCS increased the excitability of the stimulated fibres, whereas anodal tsDCS decreased it. [Color figure can be viewed at wileyonlinelibrary.com]

Analysis

Effects of the polarizing current were estimated from changes in the size and/or the latencies of potentials evoked during 5 min of consecutive polarization periods, 5 min of between‐polarization periods and 30–60 min of the postpolarization period, in comparison to control potentials (those evoked prior to the first polarization period). The areas of these potentials were assessed using software for sampling and analysis developed by E. Eide, T. Holmström and N. Pihlgren (University of Gothenburg) and were normalized with respect to the mean values of the control areas. In each experimental series, the mean control values were based on measurements of areas of 60 test and conditioned field potentials (three averages of 20 single records or six averages of 10 single records during 15 min). The measurements during each successive 5 min period were made on one or two averages of 20 or 10 single records. The normalized areas from all experiments were averaged for each testing period, and the comparisons between the periods were performed with one‐way or two‐way repeated‐measures ANOVA using the software package Statistica 12 (StatSoft Inc., Tulsa, OK, USA). When a significant effect was found (P < 0.05), the post hoc analysis for all comparisons was performed using Tukey's HSD test. Student's paired t test for paired two samples assuming equal variance was also used to estimate statistically significant differences between values at selected time periods and control values. Pearson's correlation coefficient was used to evaluate the relationship between the results.

Experimental design

Presynaptic inhibition of skin afferents was tested on dorsal horn extracellular field potentials evoked from the sural nerve, using single conditioning stimuli applied to the peroneal nerve at intensities sufficient to activate both cutaneous and group II muscle afferents. Effects of DC on presynaptic inhibition of field potentials from group Ia afferents were estimated from changes in potentials evoked by stimulation of the peroneal nerve in the motor nucleus and conditioning stimulation of either the femoral or peroneal nerves. The conditioning stimuli accordingly involved muscle afferents (in the quadriceps and sartorius nerves or the deep peroneal nerve) as well as cutaneous afferents (in the saphenous or the superficial peroneal nerves), but any changes in the field potentials from group Ia afferents were attributed to Ib afferents, which are the main source of presynaptic inhibition of these afferents (Eccles et al. 1962 c, 1964). At conditioning–testing intervals of 10–20 ms, monosynaptic components of field potentials evoked by skin afferents (within the time windows of 0.5–1.3 ms from their onset) were depressed to 64–75%, whereas those evoked by group Ia afferents in the peroneal motor nucleus (within 0.5–1.0 ms time windows) were depressed to 78–86%. The degree of the depression was as in the original studies of presynaptic inhibition at similar conditioning–testing intervals (Eccles et al. 1961, 1962 d).

Post‐activation depression was likewise tested on monosynaptically evoked extracellular field potentials of cutaneous and group II origin in the dorsal horn (Hammar et al. 2002) and on field potentials evoked by group Ia afferents in the peroneal motor nuclei. Post‐activation depression was evoked by stimulating the same nerve twice, the first time 4.5 s after the preceding stimulus, to allow the nerve to recover from effects of this stimulus, and the second time within a much shorter interval. The shorter interstimulus intervals of ∼50 ms (47–70 ms) were selected because post‐activation depression of monosynaptically evoked EPSPs recorded intracellularly in motoneurons in acute experiments in cats was found to be maximal at these intervals (Curtis & Eccles, 1960; Hultborn et al. 1996). The EPSPs induced after a train of stimuli (2T, 500 ms duration, 20 Hz) applied to the same nerve were then depressed to ∼70–80% of those evoked at ∼10 s intervals, and returned to the original size within about 5 s (see Figs 6 and 7 of Hultborn et al. 1996; Pelletier & Cicchetti, 2015; Hedegaard et al. 2015). However, we used single conditioning stimuli rather than trains, because single stimuli resulted in a comparable post‐activation depression. The depression was to 60–80% (n = 20) in the dorsal horn and to 80–90% (n = 26) in the peroneal motor nucleus.

Figure 6. Examples of effects of cathodal tsDCS on presynaptic inhibition of field potentials in the motor nucleus.

A–D, examples of field potentials recorded in the peroneal motor nucleus. In each pair of the superimposed potentials in the right part of the traces, those evoked by the test 1.8T stimuli alone were larger. Smaller potentials were evoked by the same stimuli when these were preceded by stimuli supramaximal for the group I afferents. The difference between the test and conditioned field potentials is attributed to presynaptic inhibition, and B and D show that this difference was increased after either the cathodal or anodal tsDCS; in the illustrated case, 60 min after tsDCS. Lower traces are afferent volleys recorded from the surface of the spinal cord close to the side of insertion of the microelectrode. The negativity is downwards in the microelectrode recordings and upwards in cord dorsum recordings. Vertical dotted lines indicate the time window for the measurements of the area used as a measure of the presynaptic inhibition. E and F, relationships between the areas of the earliest components of the field potentials depressed by presynaptic inhibition and the tsDCS timing. G and H, similar relationships but for the test responses alone. [Color figure can be viewed at wileyonlinelibrary.com]

Figure 7. Changes in presynaptic inhibition and post‐activation depression during and following tsDCS.

The format of the figure is as in Fig. 4 but for effects of tsDCS. A, the time course of mean effects of tsDCS (0.8–1.0 mA) on presynaptic inhibition of field potentials from the sural nerve evoked in the dorsal horn following conditioning stimulation of the peroneal nerve. Open circles close to the left ordinate indicate the control level of the presynaptic inhibition (the mean control level is indicated by the lower blue horizontal line). Filled blue circles and open diamonds indicate effects of the conditioning stimuli during and between periods of DC application as a percentage of the control during the same period, and filled diamonds indicate the postpolarization effects. Red symbols and right ordinate are for effects of anodal DC on another sample of dorsal horn field potentials. The vertical dotted line indicates the beginning of the postpolarization period. The horizontal dotted lines enclose 5–10% changes. B, as in A but for presynaptic inhibition of field potentials evoked from the group I afferents in the peroneal motor nucleus. C and D, as in A and B but for effects of DC on post‐activation depression of field potentials evoked in the dorsal horn and in the motor nucleus. Statistically significant changes in the post‐activation depression following cathodal tsDCS were revealed with one‐way repeated‐measures ANOVA (F _21,147 = 2.72, P = 0.0002) and with Student's paired t test; they are indicated by large and small asterisks, respectively (see Table 1). No effects tested with one‐way repeated‐measures ANOVA were found for either cathodal or anodal polarization in A (F _19,57 = 1.67, P = 0.07; and F _16,48 = 0.68, P = 0.80, respectively), B (F _21,147 = 0.84, P = 0.67; and F _19,76 = 0.51, P = 0.95, respectively) and C (F _21,42 = 1.19, P = 0.31; and F _17,34 = 0.81, P = 0.67, respectively) or for anodal polarization in D (F _24,120 = 1.55, P = 0.07). However, Student's paired t test revealed that third and fourth values between anodal DC and the first value after anodal DC in D were different from control values. ^* P < 0.05 and ^** P < 0.01. [Color figure can be viewed at wileyonlinelibrary.com]

In order to differentiate between the depression attributable to presynaptic inhibition, to post‐activation depression or to joint effects of the two, we used the following measures. Presynaptic inhibition was examined using heteronymous conditioning stimuli, because conditioning stimulation of the same nerve could potentially induce both post‐activation depression and presynaptic inhibition. Furthermore, presynaptic inhibition was evoked at a high repetition rate (2 Hz), resulting in a near‐maximal depression of the post‐activation depression. Post‐activation depression was, in contrast, investigated at a much lower repetition rate (see above) and at much longer interstimulus intervals, at which effects of presynaptic inhibition were negligible. In a series of preliminary experiments illustrated in Fig. 2, we ensured that hardly any depression of field potentials could be attributed to presynaptic inhibition in our experimental conditions when conditioning stimuli preceded them at intervals of about 50 ms or longer. To this end, we compared the degree of presynaptic inhibition of field potentials evoked by the second stimulus while increasing intervals between the conditioning and testing stimuli, as illustrated in Fig. 2 A. In six experiments, the mean peak amplitude of potentials evoked at 47–55 ms intervals was 98 ± 2% of the control field potentials. The range of interstimulus intervals at which presynaptic inhibition could be detected was routinely verified at the beginning of each experiment, and longer intervals (up to 70 ms) were used when needed.

Figure 2. Presynaptic inhibition of skin afferents at increasing interstimulus intervals.

A, examples of field potentials evoked by double stimuli applied with increasing interstimulus intervals at 2 Hz. The field potentials were evoked in the dorsal horn by stimulation of the peroneal nerve at three times threshold (3T). Note that the field potentials evoked by the second stimulus at 8.4–45.6 ms intervals were smaller than those evoked by the first stimulus, but the last ones returned to the original amplitude. B, the relationship between the interstimulus interval (abscissa) and the peak amplitude (ordinate) of field potentials evoked by the second stimulus illustrated in A. At 47–55 ms conditioning–testing intervals (at 2 Hz) indicated by dotted vertical lines, they reached 98 ± 2% (mean ± SEM).

Results

Effects of local cathodal and anodal polarization on presynaptic inhibition and post‐activation depression

Figure 3 illustrates the enhancement of presynaptic inhibition and post‐activation depression of an extracellular field potential evoked in the peroneal motor nucleus by 0.3 μA local cathodal polarization. The timing of the test and conditioning stimuli and the overall effects of the conditioning stimuli in control conditions are shown in Fig. 3 A–C. Expanded records in the middle and right rows show that the depression was stronger after (Fig. 3 H and I) than before the cathodal polarization (Fig. 3 E and F). As indicated by the percentages to the left of the records, 50 min after the termination of the polarization, presynaptic inhibition evoked a depression to 62% rather than 79% and post‐activation depression resulted in a depression to 61% rather than 71%. The differences (lower traces in Fig. 3 E, F, H and I) thus amounted to 17 and 10%, respectively. The mean effects of locally applied cathodal and anodal DC on all field potentials evoked from the skin and group Ia afferents in the dorsal horn and in the motor nucleus are summarized in Fig. 4.

Figure 3. Examples of changes in presynaptic inhibition and post‐activation depression following local cathodal polarization applied in the peroneal motor nucleus.

A–C, extracellular field potentials from the group Ia afferents in the peroneal nerve recorded immediately before DC application (averages of 10 single records). They were evoked by 1.6 times threshold (1.6T) stimulation of the peroneal nerve, alone (A) or preceded by 2T stimulation of this nerve to evoke post‐activation depression (at 50 ms intervals; B), or by a train of three 3T stimuli applied to the quadriceps (Q) nerve as a source of presynaptic inhibition (at 10 ms intervals; C). The two sequences of stimuli were applied every 5 s during the same 5 min period. D–F, ×20 expanded boxed parts of records in A–C. G–I, similarly expanded records of field potentials evoked by the same stimuli, but 50 min after the last period of 0.3 μA cathodal polarization. Note decreases of the control field potentials recorded before DC (to 71 and 79% by post‐activation depression and presynaptic inhibition, respectively) and more marked decreases following DC application (to 61 and 62%, respectively, with respect to the immediately preceding field potentials). In order to help to visualize these differences, the control field potentials (in grey) were overlaid upon them, and the computer‐generated difference areas are shown underneath. The negativity is downwards in the microelectrode recordings and upwards in records of afferent volleys from the surface of the spinal cord (lower traces in A–C). Vertical dotted lines indicate time windows within which the areas of the early parts of the field potentials were measured. [Color figure can be viewed at wileyonlinelibrary.com]

Figure 4. Changes in presynaptic inhibition and post‐activation depression of field potentials evoked by locally applied DC.

A, time course of mean effects of locally applied DC (0.3–0.4 μA) on presynaptic inhibition of field potentials from the sural nerve in the dorsal horn (DH) following conditioning stimulation of the peroneal nerve. Blue horizontal line and open circles close to the left ordinate indicate the control level of the presynaptic inhibition. Blue filled circles and open diamonds indicate stronger effects of the conditioning stimuli during and between periods of DC application, while the postpolarization effects are indicated by filled diamonds. Red symbols and right ordinate are for effects of anodal DC on another sample of dorsal horn field. The vertical dotted line indicates the beginning of the postpolarization period. The horizontal dotted lines enclose 5 and 10% changes in the presynaptic inhibition. B, as in A, but for presynaptic inhibition of field potentials evoked from the group I afferents in the motor nucleus (MN). C and D, as in A and B, but for effects of DC on post‐activation depression of field potentials evoked in the dorsal horn and in the motor nucleus. Statistically significant effects of cathodal and anodal polarization in A and cathodal polarization in D revealed by one‐way ANOVA (F _17,136 = 2.09, P = 0.01; F _17,153 = 2.39, P = 0.002; and F _19,152 = 2.59, P = 0.0007, respectively) are indicated by large asterisks. Those in relationship to the measurements after the first cathodal polarization in D (Tukey's HSD test) are indicated by #. Effects of cathodal and anodal polarization in B (F18,108 = 1.56, P = 0.09; F19,76 = 0.51, P = 0.95, respectively) and C (F16,112 = 1.69, P = 0.058; F17,119 = 0.82, P = 0.67, respectively) and anodal polarization in D (F24,96 = 1.62, P = 0.052) did not reach the statistical significance. Statistically significant differences with respect to control effects revealed by Student's paired t test are indicated by small asterisks (see also Table 1). ^* P < 0.05 and ^** P < 0.01; ^# P < 0.05 and ^## P < 0.01. [Color figure can be viewed at wileyonlinelibrary.com]

Presynaptic inhibition

Figure 4 A shows that presynaptic inhibition in the dorsal horn reduced the monosynaptic components of field potentials evoked from the sural nerve in control conditions to 72–74% (blue open circles and blue horizontal line, left ordinate) and that it became more effective during and following cathodal DC application, when the same conditioning stimuli decreased the tested field potentials to <70%. In contrast, anodal polarization made presynaptic inhibition less effective because it reduced its effects from 63–66% (red open circles, red horizontal line, right ordinate for another sample of the test potentials) to 70–75%. Both the enhancement and the weakening of the presynaptic inhibition developed slowly, building up during (filled circles) and between (open diamonds) the successive periods of DC application, but also after the polarization had been terminated (to the right of the dotted vertical lines).

Effects of cathodal polarization within the motor nucleus (Fig. 4 B) showed a similar trend, as the presynaptic inhibition was enhanced from the original level of 77–82% to <76–69% during DC application, although these differences were not statistically significant. Between the polarization periods, the differences were small, but presynaptic inhibition started to enhance during the postpolarization period (blue symbols). Anodal polarization applied in the motor nucleus depressed field potentials evoked from the group I afferents during DC application but hardly evoked any effects between the polarization periods or after the last DC application.

Post‐activation depression

The effects of cathodal DC on post‐activation depression were similar to or stronger than those on presynaptic inhibition. As shown in Fig. 4 C and D, the enhancement of post‐activation depression was more often by 5–10% (between the two dotted horizontal lines) than ≤5%. In contrast, overall effects of anodal polarization were either weaker or negligible (Fig. 4 C and D and Table 1A–H). Statistically significant effects were revealed by ANOVA only for the data in Fig. 4 A (cathodal and anodal) and D (cathodal; see the legend) and by Student's paired t test for cathodal but not anodal DC in Fig. 4 A, C and D and for anodal DC in Fig. 4 A (see Table 1M–P).

Table 1.

Comparison of effects of locally and trans‐spinally applied cathodal and anodal DC

		Cathodal DC			Anodal DC
	Post‐activation depression	Effects of DC (% test)	Student's paired t test	Difference (control minus DC; %)	Effects of DC (% test)	Student's paired t test	Difference (control minus DC; %)
	Motor nucleus
A.	Control for locDC	84 ± 2	—	—	85 ± 2	—	—
	Control for tsDCS	90 ± 2	—	—	86 ± 2	—	—
	H reflex (1 Hz) control†	43 ± 16	—	—	25	—	—
B.	Fifth locDC	73 ± 5**	0.009	11	86 ± 1	0.83	−1
	Fifth tsDCS	81 ± 2***	0.0004	8	86 ± 2	0.67	0
C.	Post‐locDC 10 min	78 ± 3*	0.01	6	83 ± 1	0.55	2
	Post‐tsDCS 10 min	84 ± 2	0.14	6	85 ± 1	0.43	1
	H reflex post‐tsDCS 1 min†	31 ± 10	—	12	30	—	−5
D.	Post‐locDC 25–30 min	76 ± 4*	0.01	9	85 ± 1	0.78	0
	Post‐tsDCS 25–30 min	74 ± 8**	0.003	8	84 ± 1	0.22	2
	H reflex post‐tsDCS 15 min†	26 ± 12	—	17	29	—	−4
	Dorsal horn
E.	Control for locDC	75 ± 2	—	—	71 ± 2	—	—
	Control for tsDCS	73 ± 1	—	—	63 ± 2	—	—
F.	Fifth locDC	68 ± 4*	0.05	7	71 ± 2	0.65	0
	Fifth tsDCS	66 ± 3	0.12	7	60 ± 3	0.36	3
G.	Post‐locDC 10 min	68 ± 6	0.15	7	73 ± 2	0.47	−2
	Post‐tsDCS 10 min	68 ± 4	0.36	5	59 ± 2	0.39	4
H.	Post‐locDC 25–30 min	67 ± 5*	0.03	8	73 ± 1	0.12	−2
	Post‐tsDCS 25–30 min	66 ± 4	0.30	7	61 ± 2	0.49	2

	Presynaptic inhibition	Effects of DC (% test)	Student's paired t test	Difference (control minus DC; %)	Effects of DC (% test)	Student's paired t test	Difference (control minus DC; %)
	Motor nucleus
I.	Control for locDC	77 ± 8	—	—	81 ± 4	—	—
	Control for tsDCS	85 ± 3	—	—	77 ± 7	—	—
J.	Fifth locDC	68 ± 4	0.14	9	84 ± 4	0.57	−3
	Fifth tsDCS	80 ± 5	0.11	5	79 ± 7	0.27	−2
K.	Post‐locDC 10 min	73 ± 7	0.54	4	78 ± 6	0.63	3
	Post‐tsDCS 10 min	82 ± 7	0.19	3	76 ± 6	0.67	1
L.	Post‐locDC 25–30 min	72 ± 7	0.49	5	77 ± 5	0.43	4
	Post‐tsDCS 25–30 min	81 ± 5	0.17	4	72 ± 7	0.10	5
	Dorsal horn
M.	Control for locDC	73 ± 2	—	—	65 ± 3	—	—
	Control for tsDCS	72 ± 3	—	—	77 ± 6	—	—
N.	Fifth locDC	69 ± 2*	0.03	4	69 ± 3**	0.008	−4
	Fifth tsDCS	79 ± 6	0.10	−7	66 ± 4	0.23	10
O.	Post‐locDC 10 min	68 ± 3*	0.02	5	71 ± 3*	0.02	−6
	Post‐tsDCS 10 min	66 ± 4	0.87	6	69 ± 6	0.43	8
P.	Post‐locDC 25–30 min	68 ± 3*	0.03	5	70 ± 3*	0.03	−5
	Post‐tsDCS 25–30 min	69 ± 6	0.24	3	75 ± 4	0.55	2

A, control data for effects of locally and trans‐spinally applied DC (locDC and tsDCS) on post‐activation depression of field potentials evoked in the motor nucleus in the rat and on tsDCS‐evoked changes in the human H reflex. B–D, post‐activation depression during application of DC and during two postpolarization periods in the motor nucleus. E–H, as in A–D but for DC effects on post‐activation depression in the dorsal horn. I–L, as in A–D but for DC effects on presynaptic inhibition in the motor nucleus. M–P, as in A–D but for DC effects on presynaptic inhibition in the dorsal horn. The data show the degree of the depression, the results of Student's paired t test used to compare the responses evoked before, during and after DC application and the differences between them. ^†Data on effects of tsDCS on post‐activation depression of human H reflexes are from Fig. 1 and Table 1 of Winkler et al. (2010). ^* P < 0.05, ^** P < 0.01 and ^*** P < 0.001 by Student's paired two‐sample t test, assuming equal variance.

Effects of tsDCS

Figure 5 A illustrates the set‐up used for examining the effects of tsDCS when the current was applied between two electrodes placed above the vertebral column and in contact with the abdomen, respectively. The intensity of DC to be used in these conditions was defined by monitoring effects of DC on the excitability of fibres stimulated within the dorsal horn. As illustrated in Fig. 5 B and C, tsDCS was effective at 0.2 mA, and 1.0–1.1 mA was needed for maximal effects. As a routine, therefore, we used a the submaximal tsDCS intensity of 0.8 mA.

In order to compare effects of local and trans‐spinal polarization, the effects of tsDC were tested on 11 dorsal horn field potentials and 14 field potentials evoked in the motor nuclei. Effects of cathodal tsDCS resembled those of cathodal locDCS, because tsDCS led to a similar enhancement of presynaptic inhibition and post‐activation depression of these potentials. However, anodal tsDCS has unexpectedly been found to have effects similar to rather than opposite to effects of cathodal tsDCS. Facilitation following anodal as well as cathodal tsDCS is illustrated in Fig. 6 with stronger presynaptic inhibition of field potentials evoked by group I muscle afferents in a motor nucleus. Examples of superimposed test and conditioned records in Fig. 6 A and B and in Fig. 6 C and D show that in this particular case, the presynaptic inhibition was increased from 88 to 70% after the cathodal tsDCS and from 75 to 54% after the anodal tsDCS. The time course of these changes is plotted in Fig. 6 E and F. It will be noted that the increase in the presynaptic inhibition was not associated with a decrease of the test field potentials, which was in fact increased by cathodal tsDCS (compare grey traces in Fig. 6 B with grey traces in Fig. 6 A). Figure 6 G shows also that the increase in the field potentials was as long lasting as the increase in the presynaptic inhibition.

Effects of tsDCS on the entire sample of field potentials evoked in the motor nuclei are summarized in Fig. 7 B and D and in Table 1B–D and J–L. Presynaptic inhibition and post‐activation depression of field potentials evoked in the motor nucleus appeared to be affected to a similar degree, whereas effects of tsDCS in the dorsal horn were more differentiated. Post‐activation depression was affected as in the motor nucleus (being enhanced by both cathodal and anodal DC; Fig. 7 C and Table 1F–H), whereas presynaptic inhibition showed a tendency for mixed facilitation and depression (Fig. 7 A and Table 1N–P).

Could changes in presynaptic inhibition explain changes in monosynaptic actions of skin and group I field potentials evoked by DC?

Theoretically, the decreases of field potentials evoked by local cathodal polarization could be attributable to a stronger presynaptic inhibition, whereas a weaker tonic presynaptic inhibition could underlie the increases evoked by anodal polarization. However, the results summarized in Fig. 4 A and B show that the effects of DC in the dorsal horn were in keeping with this possibility but those in the ventral horn were not. The plots show that cathodal locDC enhanced presynaptic inhibition of field potentials evoked in motor nuclei as well as in the dorsal horn, whereas it decreased field potentials evoked by skin afferents in the dorsal horn (Bolzoni & Jankowska, 2015; Jankowska et al. 2016), and effects of anodal locDC were in the opposite direction. In view of the considerable variability of these effects, a comparison of changes in a number of individual field potentials was considered to be required to increase the confidence in the discrepancy between mean DC effects on field potentials and on presynaptic inhibition of these potentials. This was performed for two time periods, i.e. during the last 5 min of locDC application and during the period of 40–50 min after it had been terminated. The results summarized in Fig. 8 A and B show that the enhancement of presynaptic inhibition of a field potential, whether of group I or cutaneous origin by cathodal DC (ordinate), was associated with larger as well as with smaller field potentials (abscissa), i.e. with data points in both lower quadrants. The enhancement of presynaptic inhibition is indicated by the positive changes (e.g. by 10% of additional depression, from 80% of the area of the control field potential before DC to 70% after DC). The same plots show that weaker presynaptic inhibition following anodal DC was most often associated with smaller field potentials irrespective of their origin, i.e. with data points in the left upper quadrants; weaker presynaptic inhibition is reflected by the negative values (e.g. by 5% weaker depression, from 80 to 85% of the control field potentials). However, only in the case of locally applied cathodal DC was there a slight correlation between changes in presynaptic inhibition and in the areas of the field potentials (Pearson's correlation coefficient r = −0.51; P = 0.03).

Figure 8. Relationships between changes in presynaptic inhibition and changes in synaptic actions of primary afferents evoked by locDC.

Relative increases or decreases in the areas of early components of field potentials evoked by test stimuli (abscissa) with respect to increases or decreases in presynaptic inhibition of these potentials during and after DC application (ordinate). The relative increases of presynaptic inhibition (within the range of 0–22%) are plotted below the horizontal dotted lines and relative decreases (within the range of 0 to −15%) above these lines. A, changes in the dorsal horn. B, changes in the peroneal motor nucleus. Data for all individual potentials were averaged as in Fig. 4, but only those for two time periods are plotted: for the last period of DC application (filled symbols) and during 40–50 min of the postpolarization period (open symbols). The plotted changes were evaluated by subtracting data for individual potentials from the control values. [Color figure can be viewed at wileyonlinelibrary.com]

Taken together, these results are not in keeping with the decrease or increase of the control field potentials by DC primarily as a function of presynaptic inhibition of these potentials. We therefore favour the possibility that DC‐evoked facilitation or depression of postsynaptic actions of primary afferents, and the modulation of these actions by presynaptic inhibition, are induced in parallel but not via the same mechanisms.

Comparison of DC effects on presynaptic inhibition and on post‐activation depression

The overall DC effects on presynaptic inhibition and on post‐activation depression presented in the previous sections revealed both similarities and differences. As summarized in Figs 4 and 7, the similarities were in the direction and in the duration of DC‐evoked changes, although the dorsal horn effects of anodal locDC were stronger on presynaptic inhibition, and the effects of cathodal locDC were stronger on post‐activation depression (Fig. 4 A and C). Whenever the similarities were found, they appeared particularly pronounced in records of the same field potentials and during the same time periods. The comparison of effects of locDC was therefore made not only using their mean values but also for individual potentials. To this end, we used 15 field potentials recorded in experiments in which presynaptic inhibition and post‐activation depression were evoked during the same 5 min periods, as outlined in Fig. 1 and illustrated in Fig. 4. The control field potentials were evoked at 5 s intervals, at which post‐activation depression following a previous stimulus was expected to be negligible. In the subsequent alternating pairs of stimuli, those used to evoke control field potentials were preceded by stimulation of the same nerve at 40–50 ms intervals, to induce post‐activation depression, or by stimulation of another nerve at 5–12 ms intervals, to induce presynaptic inhibition. In the dorsal horn, the control field potentials were evoked by stimulation of the sural nerve with conditioning stimuli applied to the sural nerve at 2T or to the peroneal nerve at 5T. In motor nuclei, the field potentials were evoked by stimulation of the peroneal nerve with conditioning stimuli applied to the quadriceps nerve. Changes evoked by the conditioning stimuli during and after DC application were then related to the degree of presynaptic inhibition and post‐activation depression found before the polarization.

All field potentials in a series were then ranked in the order of increasing areas of field potentials reduced by post‐activation depression and paired by the areas reduced by presynaptic inhibition during the same 5 min periods before, during or between and after DC application. A total of 15 such paired comparisons were made, including four for cathodal DC and five for anodal DC applied in the motor nucleus and six for anodal DC in the dorsal horn. Examples of such paired comparisons are shown in Fig. 9 A, B and C, respectively. The plots show that the variations in the areas of the individual potentials in a pair differed to some extent but occurred largely in parallel. When the potentials were reduced to the same extent by presynaptic inhibition and post‐activation depression before DC application, as in Fig. 9 A and B, no statistically significant differences were found between the areas of the potentials reduced by post‐activation depression and by presynaptic inhibition within any of the three periods (Student's paired t test for paired samples). During and after anodal polarization, two series of changes caused by post‐activation depression and presynaptic inhibition showed a similar parallelism as in Fig. 9 B and C, but in four series, one in the motor nucleus and three in the dorsal horn, the changes were parallel during DC application, but presynaptic inhibition appeared to be stronger during the postpolarization period.

Figure 9. Comparison of DC‐evoked changes in post‐activation depression and in presynaptic inhibition of the same field potentials.

A, ratios between the areas of a field potential evoked from the group I afferents in the peroneal motor nucleus, decreased by post‐activation depression (filled symbols) or by presynaptic inhibition (open symbols), and control areas. Data from the periods before, during and after cathodal locDC application are separated by vertical dotted lines. Diamonds indicate field potentials evoked during DC application, whereas circles and triangles indicate field potentials evoked during the pre‐ and postpolarization periods. Data points for potentials reduced by post‐activation depression in each of these periods were ranked in the order of the decreasing depression and paired with the data points for potentials reduced by presynaptic inhibition. B, as in A but for another field potential in the peroneal motor nucleus and anodal locDC. C, as in A but for a field potential evoked from the sural nerve in the dorsal horn and anodal locDC.

Similar changes in presynaptic inhibition and in post‐activation depression appear thus to be dominating effects of DC whether considered on individual field potentials or on their samples.

Discussion

The results of this study show that both presynaptic inhibition and post‐activation depression may be enhanced by local cathodal polarization within the terminal projection region of group Ia muscle afferents in the motor nucleus and of cutaneous and/or group II muscle afferents in the dorsal horn, and that they tend to be weakened by anodal polarization. In contrast, effects of cathodal and anodal trans‐spinally applied polarization (tsDCS) were most often similar, and both were predominantly facilitatory. The enhancement and the depression occurred during DC application as well as during the postpolarization period of at least 45–60 min.

With respect to basic aspects of effects of DC, this study has some implications for the mechanisms of DC actions in the spinal cord. Firstly, it indicates that DC effects on synaptic actions of primary afferents and on modulation of these actions by presynaptic inhibition (inferred from the changes in the extracellular field potentials) are evoked largely independently. It leads thus to the conclusion that DC‐evoked changes in postsynaptic actions may not be secondary to changes in presynaptic inhibition. Secondly, it indicates that presynaptic inhibition and post‐activation depression may share some common DC‐sensitive mechanisms. These two implications will be addressed in the last two sections of the Discussion.

Could small increases or decreases in presynaptic inhibition and in post‐activation depression by DC be functionally meaningful?

All of the previously found direct spinal effects of locally applied DC, on the fibre excitability as well as synaptic transmission, consisted of a facilitation by cathodal polarization and a depression by anodal polarization (Bolzoni & Jankowska, 2015; Jankowska et al. 2016). The enhancement of presynaptic inhibition and of post‐activation depression by cathodal locDC and the opposite effects of anodal locDC described above comply, therefore, with them. However, we have no direct way of estimating the consequences of the relatively small DC‐evoked changes found in the present study for the spinal output, even though small changes in monosynaptic Ia EPSPs evoked by post‐activation depression provide some clues in this respect. Thus, Hultborn et al. (1996) showed that post‐activation depression of EPSPs by only 15, 10 and 5% (see their Fig. 6 A, C and D) was associated with at least two to three times stronger depression of monosynaptic reflexes. Mean changes in the post‐activation depression of H reflexes following cathodal tsDCS in the study by Winkler et al. (2010) in humans by 12 and 17% when tested 1 and 15 min after tsDCS (see their Fig. 1) were likewise approximately twice as strong as 5 and 8% changes in the post‐activation depression of compound EPSPs (reflected by extracellular field potentials in the present rat preparation; Table 1C and D, column ‘differences’). It is thus likely that post‐activation depression of responses of any target cells of group I and skin afferents would be at least twice the depression of the field potentials found in the present study. There are no equivalent data on the effects of tsDCS on presynaptic inhibition in man. However, as judged by the results of the present study they may be comparable and, therefore, of use for the treatment of pathologically exaggerated or weakened spinal reflexes.

We have no ready explanation for the differences between the effects of locally and trans‐spinally applied anodal current found in the present study, nor for the differences in the effects of tsDCS in various studies in humans (Winkler et al. 2010; Lamy et al. 2012; Knikou et al. 2015). However, they may well reflect differences in the current density and its gradient within the spinal cord in conditions of tsDCS. As shown by Parazzini et al. (2014) and Song et al. (2015), the high current density during tsDCS may be expected as far ventrally as at the level of the ventral roots.

When post‐activation depression and presynaptic inhibition are enhanced by tsDCS, and the effects are sufficiently long‐lasting, they might be highly beneficial for patients with exaggerated spinal reflexes. As reviewed by Pierrot‐Deseilligny & Burke (2005), Nielsen et al. (2007), Grey et al. (2008) and Andrews et al. (2015), the exaggeration of stretch reflexes in spastic patients was first associated with abnormally weak presynaptic inhibition. However, weaker post‐activation depression of activation of motoneurons by Ia muscle spindle afferents appeared to contribute to them to a greater extent (see Pierrot‐Deseilligny & Burke, 2005). In addition, in the study of Grey et al. (2008), the maximal post‐activation depression of the H reflex in spastic patients was found to be to ∼20% weaker than in healthy subjects (see their Fig. 3, interval 1 s). The additional 12–17% enhancement of post‐activation depression of H reflexes in spastic patients by tsDCS (Table 1C and D) might thus bring it close to the normal values.

In the context of the therapeutic value of tsDCS, it might also be relevant that the analysis of the effects of tsDCS has greatly focused on monosynaptic pathways to motoneurons (see, however, Cogiamanian et al. 2012; Hubli et al. 2013; Knikou et al. 2015). However, marked post‐activation depression in pathways between cutaneous afferents and dorsal horn neurons (Hammar et al. 2002), particularly strong presynaptic inhibition of polysynaptic actions mediated by dorsal horn interneurons (Jankowska et al. 2002), and potent modulatory effects of DC in the dorsal horn (see Table 1M–P) might draw more attention to DC actions on other neuronal systems. The enhancement of presynaptic inhibition and of post‐activation depression by tsDCS might be useful to counteract spasticity attributable to hyperexcitability of interneurons in polysynaptically evoked stretch reflexes (Jankowska & Gladden, 1999; Jankowska & Hammar, 2002).

The effects of tsDCS in patients with abnormally weak spinal reflexes have been explored only with respect to locomotion‐related polysynaptic reflexes after spinal cord injuries (Hubli et al. 2013). However, recent studies indicate that spinal muscular atrophy is associated with a considerable weakening of direct input to spinal motoneurons (Mentis et al. 2011). The probability of activation of motoneurons in spinal muscular atrophy patients might thus, theoretically, be increased by counteracting the post‐activation depression and presynaptic inhibition of transmission from group Ia afferents to motoneurons. Anodal tsDCS has been found to be associated with the enhancement rather than the depression of post‐activation depression (in the present study), failed to affect post‐activation depression in humans (Lamy & Boakye, 2013), and reduced post‐activation depression in the majority of subjects in the study by Winkler et al. (2010). However, anodal locDC potently weakened presynaptic inhibition in the dorsal horn (Fig. 4 A). It might thus be worthwhile verifying the possibility of enhancing the pathologically low input to spinal motoneurons by DC.

Relationships between DC effects on synaptic actions of primary afferents and on presynaptic inhibition and post‐activation depression

Even the weakest intraspinally applied DC might have multiple effects that would depend on the region of the spinal cord. Different trajectories of presynaptic fibres and different distributions of postsynaptic neurons in motor nuclei and in the dorsal horn might therefore explain why cathodal polarization most often enhanced extracellular field potentials evoked by muscle afferents in motor nuclei, whereas it depressed dorsal horn field potentials evoked by skin afferents (Bolzoni & Jankowska, 2015; Jankowska et al. 2016). However, these opposite effects might also be attributable to different properties of the neurons or fibres or to differences in the effects of DC on presynaptic inhibition or post‐activation depression in motor nuclei and in the dorsal horn.

In previous studies, no difference has been found in the effects of locDC on the excitability of group I afferents in motor nuclei and of cutaneous afferents in the dorsal horn (Bolzoni & Jankowska, 2015; Jankowska et al. 2016). Similar effects of DC upon them therefore made it unlikely that the depression or facilitation of synaptic actions of group I and cutaneous afferents is secondary to the degree of their depolarization or hyperpolarization.

Mismatches between the effects of DC on presynaptic inhibition and on monosynaptic field potentials would be particularly relevant in this context. As illustrated in Fig. 8, we failed to find an association, unless very weak, between a decrease in the field potentials and a stronger presynaptic inhibition, or between an increase in the field potentials and a weaker presynaptic inhibition. A DC‐evoked enhancement of presynaptic inhibition occurred, at least sometimes, when the field potentials increased and decreased and would thus not be in keeping with causal relationships between the degree of presynaptic inhibition and an increase or decrease of a field potential. Accordingly, these mismatches are taken to indicate that the opposite long‐lasting DC effects on field potentials of group I and cutaneous origin are attributable to postsynaptic rather than presynaptic sites of DC actions and that they involve at least partly separate mechanisms. These mismatches also indicate that the presynaptic and postsynaptic actions are evoked independently, although in parallel. In the diagram shown in Fig. 10, this situation is indicated by separate sets of arrows, with those to the right indicating the presynaptic effects of DC on presynaptic inhibition and post‐activation depression, and those to the left indicating the effects on synaptic actions of the afferents. The mechanisms of the opposite long‐lasting DC effects on postsynaptic actions of group I and cutaneous afferents thus remain unresolved, but the search for an explanation of these effects might now be narrowed by eliminating effects secondary to presynaptic inhibition or post‐activation depression.

Figure 10. Diagram of the proposed actions of DC on presynaptic inhibition and/or post‐activation depression and on synaptic actions of afferent fibres.

For explanations see main text.

From the functional point of view, this would mean that independently DC‐evoked facilitation of synaptic actions on motoneurons might counterbalance stronger presynaptic inhibition and post‐activation depression of actions of muscle afferents in motor nuclei by the cathodal current. In contrast, postsynaptic depressive actions of cathodal current in the dorsal horn would most probably be strengthened by stronger presynaptic inhibition and post‐activation depression of synaptic actions of cutaneous afferents.

Considering the consequences of these differential actions, one should take into account that they would not be confined to artificially evoked DC effects. Effects of electrical field potentials evoked by natural stimuli (Faber & Korn, 1989) might be comparable, and it may be relevant that field potentials of cutaneous origin in the dorsal horn are particularly large.

Are common DC‐sensitive mechanisms involved in presynaptic inhibition and in post‐activation depression?

Presynaptic inhibition and post‐activation depression have been considered traditionally in their own right and, as a rule, separately. Observations of Hultborn et al. (1996) even led to the conclusion that the mechanisms of a reduction in transmitter release caused by the previous activation of these fibres during post‐activation depression are different from the ‘classical’ GABAergic presynaptic inhibition. With respect to the mechanisms of DC effects on post‐activation depression, Winkler et al. (2010) considered that ‘the increase of post‐activation depression after cathodal tsDCS might be due to… a hyperpolarization of Ia‐membrane and thus a further reduction of synaptic efficacy… while reduction of post‐activation depression following anodal tsDCS… might reflect a lasting hypopolarization in resting membrane potential of Ia afferent fibres that facilitates transmitter release…’. However, our observations link the increase of post‐activation depression to cathodal locDC and to both cathodal and anodal tsDCs, and its reduction to only anodal locDC actions.

As shown in the last section of the Results, DC effects upon presynaptic inhibition and post‐activation depression were found to be strongly linked, whether compared on overall effects of locDC or tsDCS (Figs 4 and 7) or on individual field potentials analysed during the same periods of locDC application (Fig. 9). The effects of DC upon them were similar, in particular with respect to their direction (facilitation or depression) and duration. The differences were primarily quantitative, mainly with stronger effects of anodal locDC on presynaptic inhibition and stronger effects of cathodal locDC on post‐activation depression in the dorsal horn. Consequently, these results raise the possibility of a common denominator of DC effects on post‐activation depression and presynaptic inhibition, i.e. whether they might be mediated by some common mechanisms in addition to the effects of polarization of afferent terminals. As discussed by Rudomin & Schmidt (1999), afferent fibre polarization could, for example, modulate conductance at the branching points of the afferents, or release of neuromodulators from nearby fibres or glial cells. It could thereby affect both presynaptic inhibition and post‐activation depression. This possibility is indicated in the right part of the diagram in Fig. 10. If this proves to be true, the DC‐sensitive mechanisms common for post‐activation depression and presynaptic inhibition might involve mechanisms of ‘non‐classical’ or ‘non‐traditional’ presynaptic inhibition, perhaps even PAD‐independent presynaptic inhibition.

Of particular interest in the context of the present study is that 5‐HT and the GABA_B receptor agonist baclofen affect presynaptic inhibition but not PAD, indicating that PAD is not necessarily linked with presynaptic inhibition and that presynaptic inhibition is not necessarily secondary to PAD, even if GABA may contribute to both presynaptic inhibition and to the depolarization of presynaptic terminals (see Rudomin & Schmidt, 1999, p. 8 for review of GABA‐mediated presynaptic inhibition without PAD). Recently, the importance of GABA‐mediated depolarization has been re‐assessed, and a number of alternative or additional mechanisms of PAD have been proposed (Hochman et al. 2010).

To what extent DC polarization affects these various mechanisms would be difficult to predict. Effects of DC secondary to PAD might be compatible with the close match between the enhancement of presynaptic inhibition by local cathodal DC and the increase in the excitability of the afferents by similarly applied DC. They would also be compatible with the depression of the presynaptic inhibition by local anodal DC and decreases in the excitability of the afferents. The PAD‐independent presynaptic inhibition might, in contrast, use a variety of phenomena that might also contribute to post‐activation depression. One such phenomenon might be an accumulation of potassium in the extracellular space and the ensuing fibre depolarization (Kríz et al. 1974) even in view of the evidence excluding potassium accumulation as a major contributor to PAD (Jiménez et al. 1987). The other ones might include presynaptic inhibition mediated by GABA_B receptors associated with a reduction in the duration of the presynaptic action potentials and a reduction in transmitter release via a G‐protein‐coupled mechanism (Alford & Grillner, 1991), a reduction of presynaptic Ca²⁺ inflow with a shortening of the duration of action potentials in presynaptic fibres, and a reduction in transmitter release, replicating the actions of baclofen (Curtis & Lacey, 1998), or most recently, the reported effects of spillover mechanisms proposed as one of the direct homosynaptic as well as heterosynaptic negative feedback mechanisms (Fig. 3 B of Kullmann et al. 2005; Hochman et al. 2010).

The relative impact of PAD‐independent and PAD‐dependent presynaptic inhibition has been a recurring question, and only very few tests have been found to unravel it (see Rudomin & Schmidt, 1999). One might, therefore, consider using the effects of DC and the interrelationships between presynaptic inhibition and post‐activation depression as a new tool to differentiate between presynaptic inhibition involving PAD‐associated mechanisms and PAD‐independent ones.

Additional information

Competing interests

None declared.

Author contributions

The experiments were performed at the Department of Neuroscience and Physiology, University of Gothenburg. All authors contributed to the design of the experiments, to the collection, analysis and interpretation of the data as well as to the drafting of the article. All authors approved the final version of the manuscript and agree to be accountable for all aspects of the work in ensuring that questions related to the accuracy or integrity of any part of the work are appropriately investigated and resolved. All persons designated as authors qualify for authorship, and all those who qualify for authorship are listed.

Funding

The work was supported by a grant from the National Institutes of Health (R01 NS040863) to E.J.

Translational perspective.

Trans‐spinal polarization (tsDCS) has been considered as a means to improve deficient spinal functions, but only very few of the mechanisms underlying tsDCS actions have been examined to date. The aim of the present study was to analyse one of these mechanisms, the effects of DC on two spinal modulatory systems, presynaptic inhibition and post‐activation depression. The DC effects were examined in deeply anaesthetized rats, aiming to compare any changes evoked by very weak (0.3–0.4 μA) and spatially restricted local intraspinal polarization with changes following much stronger (0.8–1.0 mA) and wider‐acting trans‐spinal polarization. Both local DC and tsDCS were found to have long‐lasting effects on presynaptic inhibition and post‐activation depression, outlasting the polarization period for at least 0.5–1 h. However, these effects differed, because local cathodal DC enhanced and anodal DC weakened presynaptic inhibition and post‐activation depression, whereas effects evoked by tsDCS were polarity independent and enhanced them irrespective of current polarity. When applied in humans, the effects of tsDCS on presynaptic inhibition and post‐activation depression might thus be beneficial in the case of exaggerated spinal reflexes, with both cathodal and anodal tsDCS counteracting them. However, they might have an adverse effect on pathologically weakened spinal activity, by an additional weakening of input to spinal motoneurons by presynaptic inhibition and/or post‐activation depression, provided that future studies will allow generalization of the results of the present study to humans.