Spinal Nociception is Facilitated during Cognitive Distraction

—The nociceptive withdrawal reﬂex (NWR) is a behavioral response to protect the body from noxious stimuli. The spatial characteristics of the stimulus modulate the reﬂex response to prevent damage to the aﬀected tissue. Interneurons in the deep dorsal horn in the spinal cord encode the relationship between stimulus characteristics and the magnitude of the NWR and are also likely integrating spatial information of the nociceptive stimulus. The aim of this study was to use the NWR to investigate whether the spinal spatial integration of a simultaneous stimulus is modulated by shifting the attention of the participant towards (attention) or away from (distraction) the stimulus. We hypothesized that the descending activity shapes the receptive ﬁelds of the spinal neurons encoding spatial integration of nociception. Twenty healthy volunteers participated in the study. Single and simultaneous stimuli were delivered through two stimulating electrodes located in the arch and on the lateral side in the sole of the foot. The NWR was quantiﬁed by electromyography from the Tibialis Anterior and Biceps Femoris muscles during baseline and active tasks (attention and distraction). During the baseline task, spatial summation of the NWR was evoked during simultaneous stimulation. During the distraction task, the NWR was signiﬁcantly larger compared to baseline, regardless of the sites being stimulated (single and simultaneous stim-uli). In contrast, the NWR recorded during the attention task did not diﬀer from baseline. These results further support that the spinal NWR pathway is under descending control which can be modulated by cognitive processes. The NWRs recorded over both proximal and distal muscles were similarly aﬀected by the tasks, suggesting that the descending control aﬀects the lower leg spinal system, with no discrimination between spinal segments. (cid:1)


INTRODUCTION
The nociceptive withdrawal reflex (NWR) is a polysynaptic reflex that serves defensive purposes in humans (Kugelberg et al., 1960;Grimby, 1963;Sandrini et al., 2005;Andersen, 2007) as well as in animals (Sherrington, 1910;Schomburg, 1990;Cleland and Bauer, 2002;Clarke and Harris, 2004). Although it was first described by Sherrington as an stereotyped flexion reaction (Sherrington, 1910), a considerable amount of evidence has been accumulated in favour of a modularly organized reflex serving a functional withdrawal reflex pattern. This organization includes control over several groups of muscles across multiple joints, leading to a net movement not necessarily characterized by joint flexion Schouenborg et al., 1994;Sandrini et al., 2005;Andersen, 2007). Indeed, the motor pattern of the NWR integrates temporal (Fuhrer, 1976;Andersen et al., 2005;Bajaj et al., 2005;Perrotta et al., 2016Perrotta et al., , 2017Sprenger et al., 2018) and spatial (Schouenborg et al., 1995;Ylioja et al., 2006;Andersen, 2007;Neziri et al., 2009;Henrich et al., 2020) characteristics of the stimulus together with proprioceptive (Serrao et al., 2014;Masse´-Alarie et al., 2019), cognitive (Bjerre et al., 2011;Jure et al., 2020) and emotional states (Rhudy and Meagher, 2000;Bartolo et al., 2013;Fragiotta et al., 2019;Lannon et al., 2021), to generate the rapid, optimal motor response to any given nociceptive input under the present posture and movement. When exposed to a suprathreshold stimulus, the recruitment of synergistic muscles moves the stimulated area away from the stimulus, coordinating the recruitment of muscle groups to keep appropriate balance and maintain the current motor activity. The spatial characteristics of the stimulus shape the reflex response to prevent damage of the exposed tissue. Among those characteristics are the location of the stimulus in the body (Grimby, 1963;Schouenborg et al., 1995;Andersen, 2007;Masse´-Alarie et al., 2019), its intensity (a more intense stimulus likely recruit a larger neuronal population), the size of the stimulated area (Chan and Dallaire, 1989;Sonnenborg et al., 2000;Andersen et al., 2001;Neziri et al., 2009;von Dincklage et al., 2013;Henrich et al., 2020), and the specific sites that are simultaneously affected by the stimulation (Henrich et al., 2020). The neural substrate behind these spatial integrative processes remains to be clarified, however, dorsal horn neurons seem to play a pivotal role (Schomburg, 1990;Coghill et al., 1991). Previous evidence has suggested that interneurons in the deep dorsal horn (DDH) encode the relationship between the intensity of the stimulus and the magnitude of the NWR (Coghill et al., 1991;Schouenborg et al., 1995;Clarke and Harris, 2004). Additionally, these neurons are likely governing the cutaneous receptive field (termed reflex receptive field, RRF) from which a NWR can be elicited .
Nociceptive stimulation of the skin generates a neural drive that converges into the dorsal horn of the spinal cord (SC). The convergent information is processed and integrated by neurons that respond selectively to nociceptive stimulation (nociceptive specific, NS), and by neurons responsive to both nociceptive and innocuous stimulation (wide dynamic range, WDR). NS neurons have small receptive fields (RF). On the other hand, RFs of WDR neurons are believed to cover a significantly larger portion of the limb (Price et al., 1978). Therefore, when two simultaneous stimuli are applied in the skin, multiple RFs of both NS and WDR neurons are likely stimulated. According to previous studies in rats (Schouenborg et al., 1992(Schouenborg et al., , 1995, WDR neurons in the DDH of the SC are likely playing a role in the encoding of the RRF of specific muscles or group of synergistic muscles, and therefore in the spatial integration of nociceptive stimuli within the reflex circuitry. An early study in rats (Weng and Schouenborg, 1996) showed that an excitatory drive onto convergent dorsal horn neurons from multiple skin sites also triggers inhibitory mechanisms. In humans, indirect evidence has shown similar filtering effects on both perception outcomes (Quevedo et al., 2017) and on the NWR (Sonnenborg et al., 2000;Henrich et al., 2020).
The afferent information reaching the dorsal horn of the spinal cord, is processed, and projected to the ventral horn and supraspinal structures. Multiple synaptic stages within the dorsal horn of the spinal cord are subject to descending modulation (Todd, 2010). The most extensively studied descending modulatory system in relation to nociception originates in the periaqueductal gray matter and relays in the rostroventral medulla (PAG-RVM) before descending through the SC (Heinricher et al., 2009). The study of the PAG-RVM as an endogenous analgesia system was motivated by early observations in animals showing inhibition of pain behavior when the PAG was electrically stimulated (Reynolds, 1969;Mayer et al., 1971;Liebeskind et al., 1973;Mayer and Liebeskind, 1974;Oliveras et al., 1975;Hayes et al., 1978). Human studies also reported inhibitory effects on nociception when the PAG was stimulated (Gol, 1967;Richardson and Akil, 1977). Although those early findings agreed on the anti-nociceptive nature of the PAG modulation, later studies confirmed that the PAG-RVM system can induce both spinal pro-and anti-nociception (Fields et al., 1983;Zhuo and Gebhart, 1992).
Among the factors that may drive descending modulation, emotional and cognitive processes have lately received considerable attention. In healthy subjects anxiety (Hubbard et al., 2011;Lannon et al., 2021) and stimuli with negative emotional valence (Rhudy et al., 2005;Roy et al., 2009;Bartolo et al., 2013;Fragiotta et al., 2019) facilitate defensive motor responses and/or self-reports of pain intensity. A cognitive manipulation of particular interest for this study is attentional shifts. In this regard, a study by Quevedo and Coghill (2007) showed that changes in attention modulated how simultaneous noxious stimuli were integrated in the perception of pain. In that study, when participants were instructed to give an overall rating for both simultaneous stimuli, spatial summation of pain was enhanced (compared to rating a single stimulus). Shifts between attention/distraction, have also shown to modulate the magnitude of the NWR (Liebermann and Defrin, 2009;Bjerre et al., 2011;Ruscheweyh et al., 2011). Particularly, Bjerre and colleagues (Bjerre et al., 2011) reported that distracting participants from the stimulus, facilitated the NWR elicited by single stimulation across different sites in the sole of the foot. On the other hand, focusing on the stimulation led to smaller NWRs. The authors interpreted those observations as an expansion/reduction of the RRF of the Tibialis Anterior (TA) muscle, produced by distraction-/attention-engaged PAG-RVM descending modulation.
Given the evidence from animal studies suggesting that WDR neurons in the DDH of the spinal cord likely encode spatial characteristics of the RRF (Schouenborg et al., 1995) one might speculate that modulation of the RRF will affect how simultaneous stimuli are integrated in the reflex pathway. The hypothesis of this study presumes that an attentional shift triggers descending modulatory drives that modulate the RRF of lower limb muscles involved in the NWR. By doing so, the integration of simultaneous stimuli is expected to be facilitated when the modulated RRF cover the two stimulation sites and inhibited when not (given the observations by Bjerre et al. 2011). Particularly, it was hypothesized that: distraction would facilitate the spatial integration of the simultaneous stimuli provoking a larger NWR magnitude when two sites in the sole of the foot within the TA RRF are simultaneously stimulated. Shifting the attention to the stimulated site was expected to shrink the RRF obstructing the integration of simultaneous stimuli -as one of the sites is no longer part of the TA RRF.

EXPERIMENTAL PROCEDURES Subjects
Interested volunteers received written and oral information about the experiment. If they agreed to participate, and met the inclusion/exclusion criteria, written informed consent was obtained. Exclusion criteria included mental disorders or disorders of the nervous or musculoskeletal system, addiction to euphoric substances, use of medicines that may affect the results of the experiment and presence of wounds on the sole of the foot. Twenty healthy subjects were included in the experiment (12 men and 8 women, mean age 26.8). The experiment was performed according to the Declaration of Helsinki and received approval from the local Ethical Committee (North Denmark Region VN-20180047).
Participants were comfortably lying on a reclined bed in a quiet room throughout the experiment. They were instructed to remain calm, relaxed and to avoid any voluntary movement during the experiment, particularly while performing EMG recordings.

EMG recordings
Double-differential EMG recordings were acquired in Tibialis Anterior (TA) and Biceps Femoris (BF) muscles (Frahm et al., 2012) in the ipsilateral leg relative to the stimulation site. Three recording electrodes (Neuroline 720, Ambu A/S, Denmark) were mounted in each muscle site following previously published recommendations on location of recording electrodes (Hermens et al., 2000). A common reference electrode (Neuroline 720, Ambu A/ S, Denmark) was placed in the ipsilateral knee over the patella bone. The EMG recordings were amplified, filtered (5-500 Hz), sampled (2 kHz), and stored for further offline analyses.

Quantification of the NWR
The NWR was elicited by applying computer-controlled electrical stimuli in the sole of the foot. The skin of the sole of the foot was prepared by a slight manual exfoliation to reduce impedance (due to the presence of thick stratum corneum layer). Two stimulating electrodes (Neuroline 700, Ambu A/S, Denmark, diameter reduced to 6 mm (Frahm et al., 2013;Henrich et al., 2020)) were mounted in the sole of the foot, medially (M) and laterally (L) (see Fig. 1). A large anodal electrode was placed on the dorsum of the foot (Pals 7.5 Â 10 cm, Axelgaard Ltd., Fallbrook, California, USA) so stimuli would be perceived in the sole of the foot. The stimulus consisted of five monophasic rectangular pulses of 1 ms delivered at 200 Hz and were delivered using an automated custommade software. Three stimulus types were used, single stimulation in M or L, and paired stimulation in both sites, M and L. The order was randomized and inter-stimulus intervals ranging between 15-30 s were used to prevent habituation (von Dincklage et al., 2013).
The NWR threshold (NWR-th) was used to set the stimulus intensity. The NWR-th was determined using an automated staircase procedure (Jensen et al., 2015a), independently for each electrode site. The staircase started with an intensity of 1 mA, increased with steps of 2 mA until a NWR was detected. Then, the intensity started decreasing until the NWR was no longer elicited. This process was repeated 2 more times with steps of 1 mA and 0.5 mA. Finally, the NWR-th was calculated as the average between the last 2 ascending and descending limits. The intensity of the stimulation was set as 150% the NWRth.
The automatic detection of the NWR was implemented in the staircase procedure described above. The criteria to decide whether a specific trial successfully elicited a NWR is based on a series of studies that attempted to standardize the NWR methodology (Rhudy and France, 2007;France et al., 2009;Jensen et al., 2015b). Specifically, the interval peak z-score (IPZ) was calculated (as in the below equation) and compared to a predefined threshold value of 12 (Rhudy and France, 2007;France et al., 2009;Jensen et al., 2015b). When the IPZ exceeds the threshold, the trial is considered as a successful NWR.
where the EMG peak is calculated in a predefined NWR window of 80-150 ms post-stimulus (Andersen, 2007). Baseline mean and standard deviation (SD) were calculated in the 70 ms pre-stimulus window.
The NWR window was defined as the interval between 80 ms and 150 ms post-stimulus as generally reported in the literature (Sandrini et al., 2005;Andersen, 2007). An electrical stimulus that is applied on the sole of the foot most likely depolarizes A and C fibers simultaneously. The difference in the conduction velocities of Ab, Ad and C fibers allow discriminating the observed NWR driven by Ad nociceptive fibers. Slow conducting C fibers can be safely excluded since its contribution to the EMG in a time window before 200 ms is highly unlikely (Hugon, 1973). However, a role of Ab fibers role cannot be completely discarded.
The NWR magnitude was quantified by calculating the root mean square value of the EMG recordings in the predefined reflex window as it is commonly reported in other studies (Terkelsen et al., 2004;Mu¨ller et al., 2016;Jure et al., 2019;Masse´-Alarie et al., 2019). Offline data processing was performed in MATLAB R2019b (Mathworks, Natick, MA, USA). Values were averaged across the five repetitions of each stimulus type for further processing.
The Detection threshold (D-th) was estimated using the same staircase procedure although stimulus detection was the assessed criteria, and increasing/ decreasing intensity steps were 1 mA, 0.5 mA and 0.1 mA.
The stimulus intensity varied between the baseline and active blocks. For the baseline block of the distraction task, it was defined as 150% of the NWRthreshold (NWR-th), while for the control block of the attention task the intensity of the stimulus was defined as 200% the detection threshold (D-th), (see Experimental protocol below).

Cognitive tasks: distraction (mStroop test) and attention (localization test)
The Distraction task was based on a modified version of a Stroop test (Stroop, 1935). To distract the participant from the stimulation, a series of slides were displayed showing different color names written in a black background with varying font colors. Slides changed at a frequency of 1 Hz and subjects were instructed to name the color of the font, ignoring the meaning of the written word. Missing/wrong words were counted as errors.
The Attention task consisted in identifying the site that was stimulated: M, L or both. During this task, participants were instructed to indicate the location of the applied stimulus in every trial. Missing/wrong localizations were counted as errors.

Experimental protocol
The experimental protocol started with a familiarization phase that consisted of a series of single and simultaneous stimulations, in random order and intensities, aiming at reducing effects of arousal and anxiety in the participant. Subjects were introduced to the cognitive tasks (mStroop and localization tests) before the final data collection began.
To control for the performance of the attention and distraction tasks, control blocks were performed. It consisted of two blocks, one for each cognitive task (Fig. 2). For distraction, the modified Stroop test (mStroop) was performed without delivering stimuli and counting the errors made when naming the color of the written word. For localization, a low intensity stimulation (C) Two different cognitive tasks were designed to shift the attention of the participant to the stimulus (attention) or away from it (distraction). In the attention task, subjects were asked to indicate the activated electrode. In the distraction task, subjects were instructed to state the color of the written word ignoring the written meaning while electrical stimulation at random intervals were delivered. mStroop: modified Stroop test.
(2xD-th, see above) was used and localization errors were quantified.
Afterwards, four stimulation blocks were conducted in random order: Attention, Distraction, and their respective Baselines (see Fig. 2). During each of those blocks, five repetitions of each stimulus type (single in M, in L and simultaneous stimulation) were delivered in random order. No specific instructions were given to the participants during baseline, other than to remain quiet and avoid any voluntary movement during the recordings. Three to five minutes resting breaks were taken between blocks.

Statistical analyses
Due to non-normal distributed data, a nonparametric Wilcoxon signed rank test was used to compare between the size of the NWR during baseline measures and the active cognitive task. For each cognitive task, a Friedman's test was used to compare between stimulus types (single in M, single in L and paired). Wilcoxon signed rank test was performed as posthoc and corrected for multiple comparison using Bonferroni correction. Wilcoxon signed rank test was used to compare the number of errors made during the cognitive tasks vs control values. Significance level was set at p = 0.05. To allow better illustration, the NWR magnitudes were displayed as logarithmic.

Baseline data
The analyses of the TA NWR during baseline conditions (Fig. 3, top panel) showed a significant effect of stimulated site (Friedman's test: v 2 (2) = 22.8, p < 0.0001). Post hoc tests showed that simultaneous stimulation elicited larger reflexes than single stimulation in the medial site (Wilcoxon signed rank test: Z = À3.6, p < 0.05), and in the lateral site (Wilcoxon signed rank test: Z = À3.9, p < 0.001). No significant difference was observed for single stimulation between M and L sites.
In contrast, analyses of the NWR in the BF muscle did not show significant effect of stimulation site (Fig. 3, bottom panel).
No significant differences were observed in the magnitude of the NWR between stimulus site (Friedman's test, p > 0.05), suggesting that the facilitation induced by the distraction task affect the entire NWR pathway irrespective of the spatial characteristics of the stimulus.
Attention task. Compared to baseline, there was no significant difference in the magnitude of the NWR regardless of the stimulated site, and recorded muscle (Figs. 4 and 5, white boxes; Wilcoxon signed rank test, p > 0.05).
Attention vs. Distraction. Direct comparison between distraction and attention showed different modulation in TA and BF. NWR recorded in TA (Fig. 4) showed facilitation during the distraction task for stimulation in M (Wilcoxon signed rank test, Z = À3.5, p < 0.001), in L (Wilcoxon signed rank test, Z = À3.0, p < 0.01), and for simultaneous stimulation (Wilcoxon signed rank test, Z = À2.7, p < 0.01). Results on BF (Fig. 5) showed a similar tendency of larger NWR during distraction, although statistical analyses revealed a nonsignificant tendency for stimulation in the lateral site (Wilcoxon signed rank test, Z = À3.0, p = 0.067), and a significant difference for simultaneous stimulation (Wilcoxon signed rank test, Z = À3.5, p < 0.01).

DISCUSSION
In this study, it was investigated whether cognitive tasks modulate the spinal integration of nociceptive stimuli through the NWR pathway, when delivering single and simultaneous stimulation of the arch and lateral side of the sole of the foot. In baseline conditions, spatial summation of TA-NWR was observed: simultaneous stimulation elicited significantly larger NWR than single stimulation.
Regarding the cognitive modulation, as hypothesized distracting the participants from the stimulus significantly facilitated the NWR. However, the degree of summation was not altered by the cognitive tasks. The NWR facilitation observed during distraction, was not different between stimulus type (M, L and Simultaneous). . . In addition, it was expected that the attention task would inhibit the NWR, an effect that could not be confirmed in this study. Finally, the NWR modulation did not differ between the Biceps Femoris and Tibialis Anterior muscles, suggesting that the descending drive affects the spinal nociceptive withdrawal reflex system with no segmental discrimination.

Top-down modulation of the NWR
Descending pathways from the brainstem support topdown modulation of spinal nociceptive pathways (Tracey and Mantyh, 2007). Animal and human studies have provided evidence for facilitatory and inhibitory influences originating from supraspinal structures (such as the PAG, dorsal reticular nucleus and ventrolateral medulla) onto spinal nociceptive circuits (see Heinricher et al., 2009 for a review). The polysynaptic spinal reflex pathway has its first synaptic connections at dorsal horn neurons exposed to descending supraspinal modulation. Specifically, PAG neurons projecting to the dorsal horn via the RVM are believed to play a key role in depressing/enhancing the nociceptive inflow at spinal level, consequently inhibiting (Zhuo and Gebhart, 1997;Danziger et al., 1998;Tracey et al., 2002;Quevedo and Coghill, 2007) or facilitating (Zhuo and Gebhart, 1997;Danziger et al., 1998;Defrin et al., 2007;Quevedo and Coghill, 2007) behavioral responses and the perception of pain.
Attempts to characterize the link between cognitive processes and descending modulation of spinal nociceptive processes are abundant in the literature. Particularly, emotions engage brain structures that likely affect the gain of the spinal nociceptive system. Fear, stress, anxiety, relaxation have been shown to modulate nociceptive responses, such as the NWR Rhudy and Meagher, 2000;Rhudy et al., 2005Rhudy et al., , 2008Roy et al., 2012;Lannon et al., 2021). Similarly, the role of cognitive processes in the modulation of pain have been extensively studied. A study by Hadjipavlou and colleagues in healthy humans has provided evidence that support the link between the PAG and many structures, more rostrally located, that are known to be active in cognitive tasks, such as prefrontal cortex, amygdala, thalamus hypothalamus and rostroventral medial medulla (Hadjipavlou et al., 2006). The presence of such anatomical connections may constitute the neural basis that support the existence of a cognitivebrainstem loop with the potential to control descending modulation in  The distraction task (compared to baseline) facilitated the NWR regardless of the stimulated site (ø: Wilcoxon signed rank test, p < 0.05). Direct comparison between distraction and attention showed that the former condition elicited larger NWR, only for simultaneous stimulation (*: Wilcoxon signed rank test, p < 0.05). Vertical axes are logarithmic for better visualization.
the nociception system. Consistent with this evidence, functional imaging studies have shown that noxiousstimulation induced activation of the spinal cord is modulated during attention tasks (Sprenger et al., 2018). However, studies manipulating participant's attention have reported discrepant results, particularly on the direction of the modulatory effect (inhibition vs facilitation). The mechanisms behind the modulation still remain elusive to characterize.
Evidence regarding the modulatory effect of attentional shifts on pain ratings generally agrees on that distracting subjects from the noxious stimuli reduces the intensity of the perceived pain (Devine and Spanos, 1990;Lautenbacher et al., 1998;Bantick et al., 2002;Tracey et al., 2002;Valet et al., 2004;Quevedo and Coghill, 2007). On the other hand, studies that reported the effect of distraction on reflex responses have shown contradictory results. An early study made by Miller and colleagues has shown inhibition of pain and reflex responses when subjects performed a demanding mental task consisting of arithmetic subtraction (Miller et al., 1979). Kiernan and colleagues reported that distracting subjects from the noxious event by the induction of hypnosis inhibit intensity of pain perception and the magnitude of the reflex (Kiernan et al., 1995). In agreement with the latter, Zachariae and colleagues also reported inhibition of the NWR when hypnotic analgesia was induced (Zachariae et al., 1998). By using a similar hypnotic methodology, Danziger and colleagues reported that eleven out of eighteen of their participants had inhibited reflex responses while in the other seven subjects the reflex was facilitated (Danziger et al., 1998). A study by Liebermann and Defrin also showed evidence supporting that cognitive tasks might induce facilitation of the NWR (Liebermann and Defrin, 2009). In the latter, when participants were unaware of the upcoming noxious stimulation, the latency of the NWR was decreased significantly (Liebermann and Defrin, 2009). In line with more recent studies (Bjerre et al., 2011;Arguissain et al., 2014), our results showed that the magnitude of the NWR was significantly facilitated during the distraction condition (Fig. 3), compared to baseline.
Contradictory results regarding modulation of the reflex might be based on inconsistencies in the methodology for manipulating attention across studies. Hypnotic analgesia studies (Miller et al., 1979;Danziger et al., 1998) may be associated with increased activation of the same brain structures (i.e.: anterior cingulate cortex (ACC), dorsolateral and orbitofrontal prefrontal cortices) involved in descending modulation of pain (Kupers et al., 2005). Studies using the Stroop test as a distracting method, similar as the one used in the present study, have shown enhanced activity in ACC and orbitofrontal areas and inhibited insula, thalamus and mid-cingulate areas involved in the codification of pain intensity (Bantick et al., 2002;Valet et al., 2004;Schumann et al., 2018).
Regarding the attention condition, no significant modulation of the NWR was observed (see Fig. 3) in this study. The attention task consisted of identifying the stimulated site. Since only three different spatial configurations were used (M, L and simultaneous), a possible explanation of the lack of modulation is that the task was not as cognitively demanding as expected. This is also reflected in the low number of errors made during the task (7%). Therefore, we speculate that since strong attentional focus was not necessary for the subject to correctly complete the task, no reflex modulation was induced.
According to our results and those discussed above, distracting the subject from a noxious stimulus may produce a differential modulation between the spinally mediated NWR and the perception of pain. Meaning a net facilitation of the NWR but inhibition of perceived pain. From a body protection perspective, it can be speculated that this modulation is advantageous for evasive behavior when a potentially dangerous stimulus is applied to the body. Indeed, the motor response is more rapid (facilitation of spinally mediated responses) while the perception of pain intensity is dampened.

Modulation of the spatial integration of nociception
The exact neural basis behind the spatial integration of nociception remains to be elucidated, challenged by the fact that direct recordings of spinal neurons are not possible in human studies. Indirect evidence has suggested that simultaneous stimuli applied in the sole of the foot in healthy humans are integrated in the spinal cord and produce an enhanced motor response from a functional perspective (Henrich et al., 2020(Henrich et al., , 2021. The spatial summation of reflexes during simultaneous stimulation in the present investigation is consistent with a spinal integration of spatially distinct nociceptive inputs. Electrophysiological studies in rats have provided evidence supporting that a set of wide dynamic range neurons (WDR), located in the DDH of the spinal cord, encode the spatial characteristics of the skin area from which a stimulus can trigger a NWR (Schouenborg et al., 1995;Morgan, 1998). The NWR circuit is likely organized in functional modules that allow the recruitment of muscles that would produce the optimal defensive behavior (e.g. flexion/extension of specific joints) while preserving balance Andersen et al., 1999;Sandrini et al., 2005). Those WDR neurons likely govern the RRFs in human subjects. For instance, TA and BF were chosen since in baseline conditions TA-RRF is restricted to the medial side of the sole of the foot while BF-RRF extends to the entire sole (Andersen et al., 2001). The extent and spatial sensitivity of the RRF can be dynamically altered by top-down modulation, as it has been shown in studies on spinal transected animals and spinal cord injured patients (Schouenborg et al., 1992;Andersen et al., 2004).
A previous study on human spinal nociception showed that distracting the subject from the noxious stimulation enlarge the RRF of the TA muscle (Bjerre et al., 2011). That dynamic enlargement of the TA-RRF was interpreted as a descending modulation affecting WDR neurons located in the DDH of the spinal cord. The cognitive activity might have rendered the sensitivity of the somatosensory system as a whole, but also shaped the extension of skin that is covered by neuronal RFs. If the spatial extent of the RRF is indeed modulated by the distraction task, it is likely that multiple stimuli applied within the expanded RRF will be integrated into a larger NWR response. However, the degree of summation when using simultaneous stimulation did not differ with single stimulation in any site. These results might suggest that the spatial integration of simultaneous stimuli per se is not under cognitive descending control, and that the modulation is inducing a facilitation of the entire NWR pathway. Moreover, the similar effects observed in TA and BF (Fig. 3) suggest that the cognitive task affect the entire spinal system with no discrimination of spinal segment.
The NWR plays a protective role in the nociceptive system. As such, the spatial characteristics of the RRF of those muscles involved in the reflex are of utmost importance, meaning that the area that is first withdrawn from the noxious stimuli correspond to the stimulated area (Schouenborg et al., 1992(Schouenborg et al., , 1994Andersen et al., 1999;Sandrini et al., 2005;Biurrun Manresa et al., 2014). Since this spatial organization of the reflex has a strong protective role, it might be advantageous from an evolutionary role, that the NWR encoding remains robust despite any cognitive modulation, which might disrupt its protective significance. This is at least partially supported by electrophysiological evidence in rats. A study has shown that DDH neurons that putatively encode the spatial characteristics of the RRF and the muscles of the NWR cannot be antidromically activated from the upper cervical spinal cord (Schouenborg et al., 1995). Although a ceiling effect that limits the summation of the simultaneous stimuli cannot be completely discarded, the most likely explanation of our results seems to be based on the functional modular organization of the NWR.
The direct comparison between the distraction and attention tasks (Figs. 4 and 5) confirmed that cognitive manipulation produces different modulatory effects onto the spinal integration of the NWR. The difference in the magnitude of the NWR between distraction and attention was significant for all stimulation sites in TA (Fig. 4), but only for simultaneous stimulation in BF. We previously argued that the localization test used to manipulate subject's attention was not demanding enough to induce a significant modulation of the NWR, compared to baseline conditions. Since the attention manipulation alone did not effectively modulate the NWR, it seems likely that the observed modulation in the direct comparison was produced by the distracting task.

LIMITATIONS
Animal studies that investigate mechanisms of spinal nociception generally use methodologies involving direct assessment of specific populations of neurons (Le Bars et al., 2001;Mogil, 2009). The direct translation of those findings to human studies is not possible and indirect methodologies to assess spinal nociception must be used. One of such methods is the quantification of the NWR as used in the present study.
To elicit the NWR, stimulation of different nature can be applied in the skin (Mørch et al., 2007). Electrical stimulation has the advantage that the stimulus onset/offset, and the stimulus intensity, can be easily controlled. Electrical stimulation of the skin is not a natural stimulus, and it bypasses the peripheral receptors and artificially depolarizes the cell membrane. Since A and C fibers are simultaneously activated, one can speculate on the partial contribution of different fiber types based on their conduction velocity (Hugon, 1973;Andersen, 2007). The definition of the NWR window in the present project is based on the Ad fiber conduction velocity. Therefore, the discussion of the results assumes that the observed NWR represents nociceptive processing particularly driven by Ad fibers. Most C fiber contribution can be excluded due to their slow conduction velocity. On the other hand, faster conducting Ab fibers cannot be completely discarded. However, it is important to note that the use of small diameter stimulating electrodes, as in the present experiment, seem to be more preferential to Ad fiber activation Frahm et al., 2013).
The present results indicate that the distraction task induces a pronociceptive state within the spinal cord that facilitates the NWR, regardless of the stimulated skin region and which joints are involved in the reflex. We hypothesized that the distracting task would expand the RRF to cover both stimulation sites (M and L), thus favoring spatial summation of simultaneous stimuli. However, the gain in the magnitude of the NWR when using simultaneous stimulus was not different to single stimulation during the distraction condition. This might be reflecting that the direction of attention to nonsomatosensory stimuli has minimal impact on spatial tuning of nociceptive processes. These results also confirm previous evidence of spatial summation within the spinal NWR circuitry. The mechanisms behind spatial integration of nociception in the spinal cord are likely playing a role in encoding/decoding spatial characteristics of pain. For instance, SSP is widely reported in the literature for simultaneous stimulation, and animal studies provided evidence of such spinal mechanisms (Coghill et al., 1993). This study provided novel evidence using the NWR to assess human spinal nociception. Based on the present results and complemented by studies on the perception of pain, it seems that the sensory integration in the spinal cord constitutes the basis of a spinal spatial filtering phenomenon that support the optimal defensive motor behavior and the corresponding perceptual experience. The NWR is then a useful objective tool to complement future research on the effect of descending cognitive modulation on spinal nociception.