Prior Methamphetamine Use Disorder History Does Not Impair Interoceptive Processing of Soft Touch in HIV Infection

Introduction: Interoception, defined as the sense of the internal state of one’s body, helps motivate goal-directed behavior. Prior work has shown that methamphetamine (METH) use disorder is associated with altered interoception, and that this may contribute to risky behavior. As people with HIV (PWH) may also experience disrupted bodily sensations (e.g., neuropathy), an important question is whether PWH with a history of METH use disorder might exhibit greater impairment of interoceptive processing. Methods: Eighty-three participants stratified by HIV infection and a past history of methamphetamine use disorder experienced a soft touch paradigm that included slow brush strokes on the left forearm and palm during blood-oxygen level-dependent functional MRI acquisition. To assess differences in interoception and reward, voxelwise analyses were constrained to the insula, a hub for the evaluation of interoceptive cues, and the striatum, which is engaged in reward processing. Results: Overall, individuals with a history of METH use disorder had an attenuated neural response to pleasant touch in both the insula and striatum. Longer abstinence was associated with greater neural response to touch in the insula, suggesting some improvement in responsivity. However, only PWH with no METH use disorder history had lower brain activation in the insula relative to non-using seronegative controls. Conclusions: Our findings suggest that while METH use disorder history and HIV infection independently disrupt the neural processes associated with interoception, PWH with METH use disorder histories do not show significant differences relative to non-using seronegative controls. These findings suggest that the effects of HIV infection and past methamphetamine use might not be additive with respect to interoceptive processing impairment.


Introduction
The use of methamphetamine (METH) among individuals at risk for HIV transmission has been called a "double epidemic" [1]; it significantly contributes to HIV transmission by enhancing risky behavior [2], and it may be associated with delays in HIV diagnosis and treatment initiation [3]. More important, its continued use in people with HIV (PWH) has been associated with reduced effectiveness of antiretroviral treatment [4] and higher plasma viral load [5,6], possibly due to poor medication adherence [5]. Independently, both HIV infection and METH can alter brain structure [7,8] and function [9][10][11][12], yet very little is known about their interaction. Some neuroimaging studies suggest there is an additive effect, whereby METH exacerbates the effects of HIV infection on brain metabolites [13] or cerebral blood flow [14]. Alternatively, structural neuroimaging studies report that METH use disorder in PWH may provide a mitigation effect [15], as it does not significantly

Imaging Procedures 2.3.1. Soft Touch Paradigm
A soft boar bristle brush (OXO International Ltd., New York, NY, USA) was administered on 4 cm long regions of skin by a trained research assistant. For a given trial, stimulation occurred on either the ventral surface of the left forearm, a region believed to contain dense mechano-receptive C-fibers, and on the palm, where these fibers are absent [21,22,52]. These regions were both pre-measured and pre-marked for consistency, and each soft brush stroke occurred at a velocity of 2 cm/s in a proximal to distal direction, standardized by an audio tone that was routed only to the research assistant's headphones [29,53,54]. This velocity has been previously shown to activate the posterior insula, and it is within the optimal range (1-10 cm/s) for C-fiber stimulation [52]. The force applied was equal to the brush's weight.
Participants performed two functional runs of the task as part of a larger functional neuroimaging battery (Figure 1). During each functional run, participants were presented with a left or rightward pointing arrow on a gray rectangular background for 3 s. Participants were asked to quickly respond by pressing either the left or right button of a button box, with the direction corresponding to the direction of the arrow. The background of the arrow would change color to indicate one of three conditions: (1) the baseline condition (gray background), in which no stimulus was expected or administered, and averaging 9 s (three consecutive arrow trials) in duration; (2) anticipation of soft touch of the left forearm (yellow background), lasting 6 s, and which indicated the participant could expect a soft touch of the forearm; and (3) anticipation of soft touch of the left palm (blue background), also lasting 6 sec, indicating the participant should expect a soft touch of the palm. Following the anticipatory periods, the soft touch condition would occur for 3 s, whereby the brush was applied to the previously indicated location for the first 2 s of the trial. Overall, anticipation and soft touch occurred twenty times for each location (palm, forearm).

Soft Touch Paradigm
A soft boar bristle brush (OXO International Ltd., New York, NY, USA) was adm istered on 4 cm long regions of skin by a trained research assistant. For a given trial, s ulation occurred on either the ventral surface of the left forearm, a region believed to tain dense mechano-receptive C-fibers, and on the palm, where these fibers are ab [21,22,52]. These regions were both pre-measured and pre-marked for consistency, each soft brush stroke occurred at a velocity of 2 cm/s in a proximal to distal direc standardized by an audio tone that was routed only to the research assistant's headpho [29,53,54]. This velocity has been previously shown to activate the posterior insula, an is within the optimal range (1-10 cm/s) for C-fiber stimulation [52]. The force applied equal to the brush's weight.
Participants performed two functional runs of the task as part of a larger functi neuroimaging battery (Figure 1). During each functional run, participants were prese with a left or rightward pointing arrow on a gray rectangular background for 3 sec. ticipants were asked to quickly respond by pressing either the left or right button button box, with the direction corresponding to the direction of the arrow. The b ground of the arrow would change color to indicate one of three conditions: (1) the b line condition (gray background), in which no stimulus was expected or administe and averaging 9 s (three consecutive arrow trials) in duration; (2) anticipation of soft to of the left forearm (yellow background), lasting 6 s, and which indicated the partici could expect a soft touch of the forearm; and (3) anticipation of soft touch of the left p (blue background), also lasting 6 sec, indicating the participant should expect a soft to of the palm. Following the anticipatory periods, the soft touch condition would occu 3 s, whereby the brush was applied to the previously indicated location for the first 2 the trial. Overall, anticipation and soft touch occurred twenty times for each loca (palm, forearm). Response accuracy and reaction time were recorded for all trials. Following the roimaging session, participants also completed a visual analog scale (VAS) questionn to independently rate pleasantness, unpleasantness, and intensity of the soft touch o forearm and palm, from "0-not at all" to "10-extremely." Response accuracy and reaction time were recorded for all trials. Following the neuroimaging session, participants also completed a visual analog scale (VAS) questionnaire to independently rate pleasantness, unpleasantness, and intensity of the soft touch of the forearm and palm, from "0-not at all" to " 10- in-plane resolution with 40 3.0 mm (2.6 mm + 0.3 mm gap) ascending interleaved axial slices) or a 3T General Electric Discovery MR 750 (Milwaukee, WI, USA) (identical parameters as above except 3.75 × 3.75 mm in-plane resolution, 40 3.0 mm ascending interleaved axial slices). To permit activation localization and spatial normalization, the following were acquired: High-resolution T1-weighted FSPGR anatomical images (Signa HDx: flip angle = 8 • , 256 × 256 matrix, 172 1 mm sagittal slices, TR = 7.77 s, TE = 2.97 ms, and 0.97 × 0.97 mm in-plane resolution; for MR 750, same as above except TR = 8.1 s, TE = 3.17 ms, 1 × 1 mm in-plane resolution). EPI-based field maps corrected susceptibilityinduced geometric distortions. The gradients system and application were not changed during the scanner upgrade, and all post processing and analysis steps were consistent across datasets. Multisite imaging studies suggested that inter-participant variance far outweighs that of site or magnet variance [55]. There were no differences in HIV or METH status, or in other demographic factors based on scanning system employed (see Table 1).

Behavioral Analysis
Response accuracy and reaction time of each button press were recorded from the onset of arrow presentation. Group level statistical analyses were performed in R (http://www.rproject.org, version 3.6.0, accessed on 26 April 2019) using a linear mixed effects (LME) model, from R's nlme package [62]. The model HIV × METH × Condition (anticipation, soft touch) × Location (palm, forearm), with Condition and Location as within-subjects factors, was used to examine both response accuracy and reaction time. The VAS predictors for "pleasantness," "unpleasantness," and "intensity" were analyzed as dependent measures using linear mixed effects to test for group differences in subjective reports. For all models, post hoc analyses were performed using R's emmeans to computer linear contrasts [63], the p-values were adjusted using the False Discovery Rate (FDR) [64], and standardized effect sizes were reported. 24

Behavioral Analysis
Response accuracy and reaction time of each button press were recorded from the onset of arrow presentation. Group level statistical analyses were performed in R (http://www.rproject.org, version 3.6.0, accessed on 26 April 2019) using a linear mixed effects (LME) model, from R's nlme package [62]. The model HIV × METH × Condition (anticipation, soft touch) × Location (palm, forearm), with Condition and Location as within-subjects factors, was used to examine both response accuracy and reaction time. The VAS predictors for "pleasantness," "unpleasantness," and "intensity" were analyzed as dependent measures using linear mixed effects to test for group differences in subjective reports. For all models, post hoc analyses were performed using R's emmeans to computer linear contrasts [63], the p-values were adjusted using the False Discovery Rate (FDR) [64], and standardized effect sizes were reported.

Behavioral Analysis
Response accuracy and reaction time of each button press were recorded from the onset of arrow presentation. Group level statistical analyses were performed in R (http://www.rproject.org, version 3.6.0, accessed on 26 April 2019) using a linear mixed effects (LME) model, from R's nlme package [62]. The model HIV × METH × Condition (anticipation, soft touch) × Location (palm, forearm), with Condition and Location as within-subjects factors, was used to examine both response accuracy and reaction time. The VAS predictors for "pleasantness," "unpleasantness," and "intensity" were analyzed as dependent measures using linear mixed effects to test for group differences in subjective reports. For all models, post hoc analyses were performed using R's emmeans to computer linear contrasts [63], the p-values were adjusted 64

Behavioral Analysis
Response accuracy and reaction time of each button press were recorded from the onset of arrow presentation. Group level statistical analyses were performed in R (http://www.rproject.org, version 3.6.0, accessed on 26 April 2019) using a linear mixed effects (LME) model, from R's nlme package [62]. The model HIV × METH × Condition (anticipation, soft touch) × Location (palm, forearm), with Condition and Location as within-subjects factors, was used to examine both response accuracy and reaction time. The VAS predictors for "pleasantness,"

Behavioral Analysis
Response accuracy and reaction time of each button press were recorded from the onset of arrow presentation. Group level statistical analyses were performed in R (http://www.rproject.org, version 3.6.0, accessed on 26 April 2019) using a linear mixed effects (LME) model, from R's nlme package [62]. The model HIV × METH × Condition (anticipation, soft touch) × Location (palm, forearm), with Condition and Location as within-subjects factors, was used to

Behavioral Analysis
Response accuracy and reaction time of each button press were recorded from the onset of arrow presentation. Group level statistical analyses were performed in R (http://www.rproject.org, version 3.6.0, accessed on 26 April 2019) using a linear mixed effects (LME) model, from R's nlme package [62]. The model HIV × METH × Condition (anticipation, soft touch) ×    Response accuracy and reaction time of each button press were recorded from the onset of arrow presentation. Group level statistical analyses were performed in R (http://www.rproject.org, version 3.6.0, accessed on 26 April 2019) using a linear mixed effects (LME) model, from R's nlme package [62]. The model HIV × METH × Condition (anticipation, soft touch) ×    Response accuracy and reaction time of each button press were recorded from the onset of arrow presentation. Group level statistical analyses were performed in R (http://www.rproject.org, version 3.6.0, accessed on 26 April 2019) using a linear mixed effects (LME) model,

Image Preprocessing
Functional images were preprocessed using Analysis of Functional NeuroImages (AFNI) [56] and FSL [57]. EPIs were slice-time corrected, motion-corrected, and aligned to high-resolution anatomical images using AFNI's align_epi_anat.py [58]. Movement parameters were visually inspected for extensive motion exceeding 3 mm, and one HIV+/METH+ dataset was subsequently excluded. Time points with isolated head movements not corrected by coregistration were censored. T1-weighted images were skull-stripped using FreeSurfer's mri_watershed [59] and registered to the MNI-152 atlas using affine transform followed by nonlinear refinement via FSL's FLIRT and FNIRT [60,61]. Functional data were aligned to standard space, resampled to 3 mm isotropic voxels, and smoothed to a 6 mm FWHM using AFNI's 3dBlurToFWHM. For each participant, AFNI's 3dDeconvolve was used to determine activation related to the soft touch paradigm. Four task regres-Viruses 2021, 13, 2476 6 of 21 sors (anticipation forearm, anticipation palm, soft touch forearm, soft touch palm) were convolved with a modified hemodynamic response function. Six motion regressors and first-, second-, and third-order polynomial trends were included as covariates of no interest. Following deconvolution, the four task-based beta regressors were converted to percent signal change.

Behavioral Analysis
Response accuracy and reaction time of each button press were recorded from the onset of arrow presentation.
Group level statistical analyses were performed in R (http://www.r-project.org, version 3.6.0, accessed on 26 April 2019) using a linear mixed effects (LME) model, from R's nlme package [62]. The model HIV × METH × Condition (anticipation, soft touch) × Location (palm, forearm), with Condition and Location as within-subjects factors, was used to examine both response accuracy and reaction time. The VAS predictors for "pleasantness," "unpleasantness," and "intensity" were analyzed as dependent measures using linear mixed effects to test for group differences in subjective reports. For all models, post hoc analyses were performed using R's emmeans to computer linear contrasts [63], the p-values were adjusted using the False Discovery Rate (FDR) [64], and standardized effect sizes were reported.

Regions of Interest
Regions of interest (ROIs) were derived from the Harvard-Oxford atlas [65]. Two bilateral ROIs were defined: an insula ROI, which contained the insula in its entirety, and a striatum ROI that included the caudate, putamen, and nucleus accumbens. These two ROIs were used as search regions for all group level fMRI analyses.

Neuroimaging Analysis
Group level statistical analyses were performed using the nlme package in R (http://www.r-project.org, version 3.6.0, accessed on 26 April 2019) to assess differences in blood oxygen level dependent (BOLD) response. Data were analyzed using an HIV × METH × Condition (anticipation, soft touch) + Location (palm, forearm) linear mixed effects approach. Location was treated as a covariate, as prior studies have not demonstrated significant interactions with location [29,53,66,67]. For all analyses, subject was nested within scanner and treated as a random effect, with HIV, METH, Condition, and Location as fixed effects. Intrinsic smoothness was estimated using the spatial autocorrelation function (acf) option in AFNI's 3dFWHMx. Minimum cluster sizes were calculated with AFNI's 3dClustSim in order to guard against false positives. For ROI analyses, a peak voxel of p < 0.001 with a cluster threshold of α < 0.025 was required for significance. This approach employs non-Gaussian models and spatial autocorrelation functions and is more robust than traditional methods [68]. A minimum cluster size of 108 µL (4 contiguous voxels) each was required for the insula and the striatum for significance. An exploratory whole brain analysis examined group differences in activation across the whole brain (peak voxel p < 0.001, cluster threshold of α < 0.05, minimum cluster size 324 µL [12 contiguous voxels]) and is presented in the supplement (Supplemental Table S2). As with the behavioral data, R's emmeans was used for post hoc analyses of significant clusters, and standardized effect sizes (ES) were reported.

Primary Robust Regression Analyses
Within-group Huber robust regressions were conducted in R to examine the relationship of clinical variables related to METH use history (age of first use, days since last use, METH use density (total quantity/total days)) within METH+ participants, and to HIV infection (illness duration in months, current CD4, nadir CD4) within HIV+ participants. Measures were natural log transformed and z-scored prior to regression. Individual regressions were performed against the mean percent signal change for anticipation palm, anticipation forearm, soft touch palm, and soft touch forearm. Significant clusters were determined within regions of interest using AFNI's 3dClustSim for small volume correction with a peak voxel of p < 0.01. Results were Bonferroni corrected for the number of measures applied to each group, two ROIs, and four conditions (α < 0.0021). Study groups did not differ on sex, handedness, or ethnicity ( Table 1). The four groups did not differ significantly with respect to age, and no HIV by METH interaction was seen. Participants with METH use disorder history (HIV−/METH+, HIV+/METH+) were significantly less educated than non-dependent participants, F(1, 79) = 5.74, p = 0.02, η 2 p = 0.068. Although there was no main effect of either METH use disorder or HIV diagnosis on WRAT-4 standard scores, the interaction of HIV with METH was significant, F(1, 79) = 6.13, p = 0.015, η 2 p = 0.072, with the HIV−/METH− and HIV+/METH+ groups scoring higher than the other two groups. The HIV+/METH− and HIV+/METH+ groups reported greater frequencies of distal symmetric polyneuropathy (p = 0.013), loss of sensation (p = 0.004), and paresthesia (p = 0.003). Groups did not differ on reports of neuropathic pain (p = 0.11) or dysesthesia (p = 0.11). There was a main effect of METH use disorder history on the BDI-II, such that participants with METH use disorder history scored higher than those with no prior METH use disorder history, F(1, 76) = 18.69, p < 0.001, η 2 p = 0.20. Individuals with METH use histories tended to have higher frequencies of most substance use disorders, as well as lifetime antisocial personality disorder, whereas all but the HIV−/METH− group had a higher frequency of lifetime major depressive disorder (Supplemental Table S1).

VAS Scales
There were no significant effects of HIV status, METH status, or their interaction for VAS ratings in relation to pleasantness or unpleasantness of soft touch to either the palm or forearm, or of the intensity of soft touch to the palm, ps > 0.06. However, there was a main effect of HIV on the VAS ratings in relation to intensity of soft touch to the forearm, F(1, 79) = 5.02, p = 0.028, ES = 1.93, such that PWH rated touch as more intense than HIV− participants.

Behavioral Analyses
There were no significant main effects of HIV diagnosis, METH use disorder history, Condition, or Location with respect to correct responses, nor were any interactions significant (all ps > 0.055). There was a main effect of Condition, F(1, 237) = 25.77, p < 0.001, ES = 4.00, with participants responding more slowly during soft touch receipt than during anticipation. There was also an HIV × METH × Condition interaction, F(1, 237) = 6.05, p = 0.016. Post hoc analyses revealed that the HIV−/METH− had slower reaction times during soft touch receipt relative to anticipation, t(237) = 2.91, p = 0.023, ES = 3.77. This relationship was also seen for the HIV+/METH+ group, t(237) = 4.67, p < 0.001, ES = 8.00. No other significant main effects or interactions were detected.

Main Effect of Condition
There was a main effect of Condition across all participants in the bilateral insula and striatum in their entirety, with an increased BOLD response to soft touch receipt relative to anticipation ( Table 2).

HIV × Condition Interaction
An HIV × Condition interaction was detected within the left insula; however, post hoc analyses suggested that participants, regardless of HIV status, showed an elevated response to soft touch receipt relative to anticipation (Table 2, Figure 2).

METH × Condition Interaction
There were several clusters, predominately within the left insula, left dorsal caudate, left putamen, and left ventral striatum, demonstrating a METH × Condition interaction. In all clusters, post hoc analyses demonstrated that both METH− and METH+ individuals showed a greater response to soft touch receipt relative to touch anticipation (Table 2, Figure 3). However, clusters within the left posterior insula, the right anterior insula, and within the left dorsal caudate and posterior putamen additionally revealed that the BOLD response to soft touch receipt in METH+ individuals was attenuated relative to METH− individuals (Figure 3). METH+ individuals also had relatively greater BOLD response than METH− individuals to soft touch anticipation within the right anterior insula.

METH × Condition Interaction
There were several clusters, predominately within the left insula, left dorsal caudate, left putamen, and left ventral striatum, demonstrating a METH × Condition interaction. In all clusters, post hoc analyses demonstrated that both METH− and METH+ individuals showed a greater response to soft touch receipt relative to touch anticipation (Table 2, Figure 3). However, clusters within the left posterior insula, the right anterior insula, and within the left dorsal caudate and posterior putamen additionally revealed that the BOLD response to soft touch receipt in METH+ individuals was attenuated relative to METH− individuals (Figure 3). METH+ individuals also had relatively greater BOLD response than METH− individuals to soft touch anticipation within the right anterior insula. ses 2021, 13, x 12 of 22 Within the left posterior insula and right anterior insula, both the METH− and METH+ groups demonstrated a greater BOLD response to touch receipt relative to anticipation. Within the right anterior insula, the METH+ group showed a greater BOLD response to anticipation relative to the METH− group; in contrast, the METH− group demonstrated greater BOLD response to touch receipt relative to the METH+ group. (B) Across all three striatal clusters, both METH− and METH+ groups showed greater BOLD response to touch receipt relative to anticipation, and the METH− group had greater BOLD response to touch receipt relative to METH+. %SC: percent signal change. * p < 0.05; ** p < 0.01; *** p < 0.001.

HIV × METH × Condition
Three clusters showing significant HIV × METH × Condition interactions were identified (Table 2, Figure 4). Within the bilateral anterior insula and the left dorsal middle insula, the HIV−/METH−, HIV+/METH−, and HIV+/METH+ cohorts demonstrated a stronger BOLD response to touch relative to anticipation. In contrast, the HIV−/METH+ group demonstrated this effect only within the left anterior insula, while within the right anterior insula the HIV−/METH+ group showed a greater BOLD response to anticipation relative to touch. Differences stratified by either HIV serostatus or METH use disorder history were also identified for both anticipation and receipt. Within HIV seronegative groups, METH+ exhibited greater BOLD responses to anticipation within the left dorsal middle insula, while METH− responded more strongly to touch receipt within the bilateral anterior insula and left dorsal middle insula. For METH+ groups, individuals who groups demonstrated a greater BOLD response to touch receipt relative to anticipation. Within the right anterior insula, the METH+ group showed a greater BOLD response to anticipation relative to the METH− group; in contrast, the METH− group demonstrated greater BOLD response to touch receipt relative to the METH+ group. (B) Across all three striatal clusters, both METH− and METH+ groups showed greater BOLD response to touch receipt relative to anticipation, and the METH− group had greater BOLD response to touch receipt relative to METH+. %SC: percent signal change. * p < 0.05; ** p < 0.01; *** p < 0.001.

HIV × METH × Condition
Three clusters showing significant HIV × METH × Condition interactions were identified (Table 2, Figure 4). Within the bilateral anterior insula and the left dorsal middle insula, the HIV−/METH−, HIV+/METH−, and HIV+/METH+ cohorts demonstrated a stronger BOLD response to touch relative to anticipation. In contrast, the HIV−/METH+ group demonstrated this effect only within the left anterior insula, while within the right anterior insula the HIV−/METH+ group showed a greater BOLD response to anticipation relative to touch. Differences stratified by either HIV serostatus or METH use disorder history were also identified for both anticipation and receipt. Within HIV seronegative groups, METH+ exhibited greater BOLD responses to anticipation within the left dorsal middle insula, while METH− responded more strongly to touch receipt within the bilateral anterior insula and left dorsal middle insula. For METH+ groups, individuals who were seronegative for HIV responded more strongly to anticipation within the right anterior insula as well as the left dorsal middle insula, whereas PWH showed a greater BOLD response to touch receipt within the right anterior insula and left middle dorsal insula.
Finally, within METH− groups, HIV seronegative individuals exhibited greater BOLD response to pleasant touch within the left anterior insula relation to HIV seropositive individuals. There were no significant findings for the HIV × METH interaction, nor were there significant main effects of either HIV or METH. were seronegative for HIV responded more strongly to anticipation within the right anterior insula as well as the left dorsal middle insula, whereas PWH showed a greater BOLD response to touch receipt within the right anterior insula and left middle dorsal insula. Finally, within METH− groups, HIV seronegative individuals exhibited greater BOLD response to pleasant touch within the left anterior insula relation to HIV seropositive individuals. There were no significant findings for the HIV × METH interaction, nor were there significant main effects of either HIV or METH.

PWH
Huber robust regression suggested that lower CD4 nadir was associated with lower BOLD response within the right anterior insula, extending into the claustrum, to the anticipation of soft touch of the palm, t(34) = 1.72 ( Figure 5A). Similarly, lower CD4 nadir was associated with a lower BOLD response within the left putamen to the anticipation of soft touch of the forearm, t(34) = 1.42.

Individuals with a History of METH Use Disorder
Longer abstinence (as measured by days since last use) was associated with a higher BOLD response the left posterior insula to pleasant touch of the palm, t(34) = 7.22 ( Figure  5B). A separate cluster within the left anterior insula also suggested that longer abstinence was associated with a higher BOLD response to pleasant touch of the forearm, t(34) = 4.25. Huber robust regression suggested that lower CD4 nadir was associated with lower BOLD response within the right anterior insula, extending into the claustrum, to the anticipation of soft touch of the palm, t(34) = 1.72 ( Figure 5A). Similarly, lower CD4 nadir was associated with a lower BOLD response within the left putamen to the anticipation of soft touch of the forearm, t(34) = 1.42.

Individuals with a History of METH Use Disorder
Longer abstinence (as measured by days since last use) was associated with a higher BOLD response the left posterior insula to pleasant touch of the palm, t(34) = 7.22 ( Figure 5B). A separate cluster within the left anterior insula also suggested that longer abstinence was associated with a higher BOLD response to pleasant touch of the forearm, t(34) = 4.25.

Discussion
The present study suggests that PWH with no METH use disorder history exhibit altered neural signals within the insula, a region involved in predicting and interpreting pleasant interoceptive stimuli in comparison to HIV−/METH− individuals. However, METH use disorder history in PWH did not appear to introduce additional interoceptive or reward processing impairment relative to PWH with no METH use disorder history. Rather, PWH with METH use disorder history were not statistically different in terms of touch anticipation and receipt relative to non-using seronegative controls, suggesting that

Discussion
The present study suggests that PWH with no METH use disorder history exhibit altered neural signals within the insula, a region involved in predicting and interpreting pleasant interoceptive stimuli in comparison to HIV−/METH− individuals. However, METH use disorder history in PWH did not appear to introduce additional interoceptive or reward processing impairment relative to PWH with no METH use disorder history. Rather, PWH with METH use disorder history were not statistically different in terms of touch anticipation and receipt relative to non-using seronegative controls, suggesting that the interaction of HIV infection with METH is not additive. Taken together, these findings support disrupted interoceptive processing, but additionally suggest this effect is minimized in PWH with histories of METH use disorder.

Significance of Findings
Overall, we did not find significant neural activation differences related solely to HIV status that were independent of METH use disorder history. Rather, for individuals without METH use disorder histories, PWH exhibited reduced neural activation to soft touch receipt in the left anterior insula relative to seronegative controls. Neuroanatomical studies have reported reduced cortical volume in PWH relative to seronegative controls in multiple regions, including the insula [69] and is further exacerbated in association with HIV-associated distal neuropathic pain in PWH [70][71][72]. Others have reported reduced intrinsic resting state activity within the insula in PWH, [73] and, more broadly, reduced functional connectivity between the insula and other regions involved in the salience network [74][75][76][77]. More recent work has reported increased right anterior insula activation to expectation of pain offset PWH with distal neuropathic pain relative to PWH without this condition [78], suggesting abnormal processing of pain relief in those with chronic pain. Our own work has reported greater activation in the anterior insula to risky decisions in PWH relative to HIV seronegative individuals [9]; however, this could have been due to increased activation in the caudate nucleus, anterior cingulate, and dorsolateral prefrontal cortex, as these regions are reciprocally connected with the anterior insula [79]. An attenuated response to soft touch could reflect changes in interoceptive awareness. Although PWH, regardless of METH history, reported similar levels of pleasantness relative to seronegative controls, they also reported that pleasant touch felt more intense. Our findings therefore suggest that there could be changes in the interpretation of affective touch that predate neurological symptoms, as the majority of our participants did not meet the threshold for distal polyneuropathy. Additional research is needed to better address this question.
We did detect a two-way interaction of METH use disorder history with Condition, whereby METH+ individuals, regardless of HIV status, demonstrated an attenuated response to soft touch receipt within the left posterior insula, right anterior insula, left dorsal caudate, and left posterior putamen relative to non-using individuals. METH has consistently been linked to deficits in striatal functioning to reward processing and risk behavior [12,80,81], as well as to interoceptive deficits in the insula [29,82,83]. Moreover, others have reported reduced insula activation during other tasks, including cognitive control [84] and decision making [85]. Our findings are consistent with the broader literature of attenuated response to pleasant stimuli in the striatum, supporting a general thesis that METH is associated with an impaired ability to process reward, along with an attenuated insula response that is indicative of an impaired ability to predict the bodily experience to exteroceptive stimuli.
Our results also demonstrated that prior METH use disorder history may interact with HIV serostatus, suggesting a possible synergistic effect on BOLD activation within the insula. The HIV−/METH+ group demonstrated reduced BOLD response to pleasant touch in the bilateral anterior insula and left dorsomedial insula. A similar relationship was found for METH− individuals, as HIV infection was associated with a lower BOLD response in the left anterior insula. However, for individuals with METH use disorder histories, PWH had greater responses to pleasant touch in the right anterior insula relative to seronegative individuals. Moreover, the HIV+/METH+ group was not statistically different from the HIV−/METH− group. Prior work examining brain morphometry has reported smaller cortical and subcortical volume and thinner cortex in PWH relative to seronegative controls, and modest reductions in cortical thickness in individuals with prior METH history, yet no significant differences in their interaction [15,16]. Similarly, others have reported nonsignificant interactions of METH and HIV infection in brain metabolites [13,86] or cerebral blood flow [14] although it has been suggested that effects of HIV and METH could be additive [13,14].
In contrast, others have reported that while main effects of HIV or METH during a complex motor task were associated with attenuated BOLD responses in the striatum and insula, the BOLD activation in the HIV+/METH+ group was more similar to HIV−/METH− [17]. In seronegative individuals, active METH use increases extracellular dopamine due to the reduction of dopaminergic transporters, increases glutamate levels, and activates inflammatory pathways [30]. HIV infection has also been associated with elevated levels of glutamate [87], impaired dopaminergic functioning [88], and inflammation. Little is known about the interaction of METH use disorder with HIV infection on inflammation, although evidence suggests it could be exacerbated, as METH-using PWH show greater cognitive impairment [89]. Both HIV infection and METH use affect the peripheral immune system, typically by elevating inflammatory cytokines and chemokines [90,91], and the comorbid use of METH in the context of HIV may additionally impact immune system homeostasis. Moreover, dopamine also regulates immune functioning, and immune systems can subsequently influence dopamine signaling [92], thereby modulating tactile perception and perceived intensity [93]. However, prolonged abstinence from METH in seronegative individuals has also been shown to largely normalize glutamate levels [94], as well as promote significant recovery of dopaminergic transporters in the brain [95]. As METH abstinence has been linked to lower viral loads in PWH [96], this could suggest some reduction of excess glutamate, which is associated with inflammation, along with less dysregulation of dopaminergic systems. For the purpose of pleasant touch, it is possible that METH abstinence in PWH might lead to modest improvements in dopamine and glutamate function, allowing for neural activity that appears more similar to non-using seronegative controls. Others have suggested that improvements related to prolonged abstinence from METH use are specific to younger PWH, as the combination of prior neurotoxicity coupled with the aging process might be especially impactful [97]. More comprehensive studies are needed to address this question.
Within all PWH, there was a modest association of lower CD4 nadir with lower BOLD responses to soft touch anticipation in the right anterior insula and left putamen. Other studies have linked lower nadir CD4 to altered risk-related neural processing [9,98]. Lower CD4 nadir has also been linked to cortical thinning and reduced cortical gray matter volume [16,[99][100][101], although others have reported no associations [102,103]. Our findings provide further support for associations between historical immunocompromise and vulnerability of HIV-related neural injury [104], including for interoception. Among individuals with METH use disorder histories, longer abstinence was associated with higher activation in the insula to receipt of pleasant touch. Other studies of interoceptive processing have not found this association; however, those studies were in recently abstinent individuals [29], suggesting that longer periods of abstinence could lead to improvements in interoceptive neural processing.

Limitations
This study has several limitations. It is not possible to determine whether findings are related to premorbid traits or whether findings are a consequence of METH exposure or presence of HIV infection. It is also possible that the initiation of METH use relative to HIV seroconversion might play a role [105]. Although our analyses were focused on METH use disorder history in the context of HIV infection, we did not exclude participants for marijuana, alcohol, or nicotine use, as these substances often overlap with METH use. It is possible that MRI eligibility and willingness to participate may have induced sampling bias due to a relatively healthier study population. This study also assessed the anticipatory response to predictable and certain events, and it's possible the level of predictability may have influenced the anticipatory response. The inclusion of a continuous performance task and an anticipatory phase may also have mitigated our ability to detect meaningful differences related to C-fiber stimulation. Attention [106] and other contextual information, such as visual stimuli or one's internal motivational state, [107][108][109] can modulate one's sensory awareness. Finally, smaller sample size, particularly for the HIV+/METH+ group, could mean our results are not generalizable, and larger studies are needed.

Conclusions
The present study examined the independent and combined effects of HIV infection and methamphetamine use disorder history on pleasant touch, an interoceptive process. We found that while METH use disorder history, regardless of HIV status, attenuated insular and striatal response to pleasant touch, a reduced response to pleasant touch in PWH was only seen in the insula for the non-using cohort. Finally, HIV+/METH+ individuals did not show statistically significant differences in neural activation from METH− seronegative controls. These findings suggest that METH and HIV may differentially affect interoceptive processing.
Supplementary Materials: The following are available online at https://www.mdpi.com/article/10 .3390/v13122476/s1. Table S1: Statistical comparisons were by means of fisher-exact tests for equality of proportions; Table S2: Significant BOLD activation for the Location (Palm, Forearm) + HIV × METH × Condition (Anticipation, Receipt) linear mixed effects whole brain analysis.
Author Contributions: A.B.-G.: conceptualization, data curation, formal analysis, funding acquisition, study performance, writing-original draft, writing-review and editing; R.J.E.: formal analysis, funding acquisition, writing-review and editing; S.F.T.: writing-review and editing; M.P.P.: conceptualization, funding acquisition, study design, writing-review and editing; I.G.: funding acquisition, study design, study performance, writing-review and editing. All authors have read and agreed to the published version of the manuscript.
Funding: Supported by a grant from the National Institute of Health, (P50DA026306). The funding source had no additional input in the study design, data collection, data analysis, data interpretation, writing of the report, or in the decision to submit the article for publication.

Institutional Review Board Statement:
The study was conducted according to the guidelines of the Declaration of Helsinki and approved by the Institutional Review Board at University of California, San Diego.

Informed Consent Statement: Subjects gave written informed consent for inclusion in the study.
Data Availability Statement: Data will be made available on request.