A randomized proof-of-mechanism trial of TNF antagonism for motivational anhedonia and related corticostriatal circuitry in depressed patients with high inflammation

Chronic, low-grade inflammation has been associated with motivational deficits in patients with major depression (MD). In turn, impaired motivation has been linked to poor quality of life across psychiatric disorders. We thus determined effects of the anti-inflammatory drug infliximab–a potent tumor necrosis factor (TNF) antagonist–on behavioral and neural measures of motivation in 42 medically stable, unmedicated MD patients with a C-reactive protein > 3mg/L. All patients underwent a double-blind, placebo-controlled, single-dose, randomized clinical trial with infliximab (5mg/kg) versus placebo. Behavioral performance on an effort-based decision-making task, self-report questionnaires, and neural responses during event-related functional magnetic resonance imaging were assessed at baseline and 2 weeks following infusion. We found that relative to placebo, patients receiving infliximab were more willing to expend effort for rewards. Moreover, increase in effortful choices was associated with reduced TNF signaling as indexed by decreased soluble TNF receptor type 2 (sTNFR2). Changes in effort-based decision-making and sTNFR2 were also associated with changes in task-related activity in a network of brain areas, including dmPFC, ventral striatum, and putamen, as well as the functional connectivity between these regions. Changes in sTNFR2 also mediated the relationships between drug condition and behavioral and neuroimaging measures. Finally, changes in self-reported anhedonia symptoms and effort-discounting behavior were associated with greater responses of an independently validated whole-brain predictive model (aka “neural signature”) sensitive to monetary rewards. Taken together, these data support the use of anti-inflammatory treatment to improve effort-based decision-making and associated brain circuitry in depressed patients with high inflammation.


INTRODUCTION
De cits in reward motivation-often referred to as "motivational anhedonia" [1][2][3] -are a common feature of mood and other psychiatric disorders and are strongly associated with impaired quality of life 4 .Prior human and laboratory animal studies have repeatedly linked motivational anhedonia to dysfunction of corticostriatal reward networks [5][6][7] that support effortful behavior 8-10 and include dopamine (DA)-rich areas such as the striatum as well as brain regions encompassing the dorsomedial prefrontal cortex including the dorsal anterior cingulate cortex and surrounding paracingulate and pre-supplementary motor areas (herein referred to collectively as "dmPFC").Human functional imaging studies as well as animal lesion studies have repeatedly implicated the dmPFC as a critical hub for effort-based decisionmaking [11][12][13][14][15] , which appears to encode multiple decision-variables related to effort cost 16 , choice di culty 17 and effort-related expectation violation 11,18 .Taken together, the dmPFC and striatum have been found to encode distinct decision-variables related to effort-based choice and appear to be causally involved in the willingness to expend effort for rewards.
A substantial body of evidence suggests that in ammation may have a signi cant impact on brain regions involved in effort-based motivation in patients with major depression (MD).Elevations in peripheral in ammatory cytokines are often present in MD patients [19][20][21][22] and have been reliably associated with symptoms of anhedonia 23 .Moreover, in ammatory stimuli have been shown to exert direct effects on dmPFC activity and striatal DA availability in laboratory animals and humans.In the case of dmPFC, prior functional magnetic resonance imaging (fMRI) and magnetic resonance spectroscopy studies have found that administration of interferon (IFN)-alpha, typhoid vaccination, or endotoxin leads to decreased task-based activation [24][25][26] as well as increased glutamate 27 .Similarly for striatal DA, studies in rodents, non-human primates and humans have found that the cytokines interleukin (IL)-1 beta, IL-6, and IFN-alpha lead to substantial decreases in DA [28][29][30][31] .Functional neuroimaging studies have also revealed attenuated activity in striatum during DA-sensitive processing, such as during reward anticipation or reward prediction errors following immune activation [32][33][34][35] .
These data suggest that effects of in ammatory signaling on corticostriatal circuitry may represent a speci c pathophysiological substrate for motivational de cits in MD.Importantly, only ~ 30% of MD patients exhibit increased in ammation, possibly representing an "in ammatory subtype" for motivational de cits 21 .Consistent with this notion, evidence suggests that targeted DAergic or antiin ammatory therapies exhibit selective bene t for symptoms of anhedonia in MD patients exhibiting high-but not low-in ammation 36,37 .To date, however, no study has tested the hypothesis that inhibition of in ammation can reverse objective measures of motivational anhedonia and related mesolimbic circuitry in depressed patients with high in ammation.Randomized participants were free from all psychotropic and anti-in ammatory medications for at least 4 weeks and were without evidence of chronic infection, autoimmune or in ammatory disorders, or unstable medical illnesses as determined by medical history, physical exam and laboratory testing (full details of eligibility criteria and assessment are included in the Supplemental Materials).No patient was removed from psychotropic treatment for the purposes of the study.Of the randomized participants, 38 had available self-report measures, in ammatory markers, and behavioral data from the EBDM task, and 37 had available neuroimaging data.Two additional participants exhibited signi cant motion (3mm-6mm) during task-based fMRI; one additional subject exhibited a low calibration response at 14 days.These subjects were included in analyses, but sensitivity analyses were performed to assess their impact on reported results (see Supplemental Materials).Finally, two participants had an insu cient number of trials for division into training and test datasets, and were therefore excluded from the multivariate analysis.

METHODS
Demographic and clinical data of randomized participants are presented in Table 1 and a full consort diagram is provided in Supplemental Figure S1.Written informed consent was obtained from all participants.The Emory Institutional Review Board granted study approval (IRB00087941).There were no serious adverse events associated with this study.A full list of adverse events is included in the Supplemental Materials.The study utilized a randomized, placebo-controlled, clinical trial design to examine the effects of a single-dose of the TNF antagonist in iximab on behavioral and brain measures of motivational anhedonia.Baseline blood, behavioral, self-report and neuroimaging assessments were followed by an infusion of either a 5mg/kg of in iximab or saline (placebo), administered over ~ 40 minutes from saline bags matched for color and consistency.Blood samples and behavioral and self-report measures were repeated at 3 and 14 days, and neuroimaging assessments including resting-state fMRI and an effortbased decision-making task 11 were repeated at 14 days (see Fig. 1A).All study personnel were blinded to group assignment.The randomization and blind key were tracked by the Emory Investigational Drug Service.

Effort Based Decision-Making Task (EBDM)
The EBDM task is an fMRI-adapted effort-based decision-making task 11,39 that measures neural responses to effort and reward magnitude.During each trial, participants were given the choice between High Effort and No Effort options.The High Effort option requires more effort (as measured by button presses) than the No Effort option.The reward obtained from the No Effort option was always $1 while rewards from the High Effort option varied between $1.00 and $5.75.The magnitude of effort required in the High Effort option consisted of 20%, 50%, 80%, and 100% of the participant's maximal effort.The task shows good internal consistency, with a split-half reliability r = .94.Participants made choices in the scanner with the effort completed as soon as the scanning session was concluded.As such, the task can be viewed as measuring the choice to commit to effort expenditure for reward in the near future (see Supplemental Materials).
Clinical Assessments: Measures of motivational anhedonia included the reduced motivation (RM) subscale of the Multidimensional Fatigue Inventory (MFI 40 ), a composite of items from the Motivation and Pleasure Scale-Self Report focused on effort 41 , and the anhedonia subscale from the Inventory of Depressive Symptomatology-Self Reported (IDS-SR 42 ).Scales were collected prior to infusion at either screening, baseline MRI or infusion visits, and then at three days and 14 days post-infusion.These scales were pre-registered (https://osf.io/r6m49/)as clinical measures of motivational anhedonia given prior associations in in ammation and striatal function 31,43 .

Biological Assays
Whole blood was collected into EDTA-containing vacutainer tubes through indwelling catheters after 30 minutes of rest to limit effects of stress.Plasma was isolated and stored at -80 until batched assay.
Customized Fluorokine MAP Multiplex Human Biomarker Panels (R&D Systems, Minneapolis, MN) were used to measure plasma soluble tumor necrosis factor receptor 2 (sTNFR2) and other in ammatory markers (see Supplemental Materials).Inter-and intra-assay coe cients of variation were reliably less than 10%, and no values were below limits of detection.Plasma CRP was measured using a high sensitivity turbidimetric assay in the CLIA-certi ed Emory Medical Laboratory.
Neuroimaging Data Acquistion: Functional and structural neuroimaging data were acquired on a Siemens 3T Tim Trio using a 32-channel phased-array head coil.Trial presentations were synchronized to initial volume acquisition.Functional (T2* weighted) images were aquired using a multiband sequence with the following sequence parameters: 3-mm 3 isotropic voxels, repetition time (TR) = 1.0 s, echo time (TE) = 30 ms, ip angle (FA) = 65°, 52 interleaved axial slices, with slice orientation tilted 18° relative to the AC/PC plane to improve the temporal signal-to-noise ratio (tSNR) and minimize signal dropout of ventromedial prefrontal cortex.At the start of the imaging session, a high-resolution structural volume was also collected, with the following sequence parameters: 2-mm × 1-mm × 1-mm voxels, TR = 1.9 s, TE = 2.27 ms, FA = 9°.

Power analysis and Impact of the Covid-19 Pandemic
Data in this study were drawn from a clinical trial (NCT03006393) focused on examining the effects of a single-dose of in iximab on brain function.The study was actively recruiting at the outset of the COVID-19 pandemic, at which point the investigators deemed continuation of the study to be unsafe, given the immune-suppressive effects of in iximab.Consequently, we were unable to achieve our originally proposed recruitment target of n = 80.As a result, the current study retains adequate power to detect large effect-sizes, but does not have adequate power for small or medium effect sizes, which may have increased type II error for some analyses.We note that prior pharmacologic studies of effort-based decision-making tasks have suggested large effect sizes (d > 1.0) 44,45 , though these studies did not use in iximab.
Pre-registration: A pre-registered analysis plan that details hypotheses, key dependent variables, and primary methods can be found at https://osf.io/r6m49/.The current paper focuses on a subset of these data.Speci cally, given the focus of the current paper on effects of in iximab on motivational circuitry, we used the secondary behavioral endpoint due to its temporal proximity to change in neuroimaging.Analyses not pre-registered are designated as "exploratory".Table S1 in the Supplemental Materials summarizes our primary, secondary and exploratory analyses.

General Statistical Methods
Multiple linear regression was used to examine associations between change from baseline to endpoint across choice data, clinical measures, sTNFR2 and fMRI data extracted from regions-of-interest (ROIs).
To assess the effects of in iximab on in ammatory markers, multivariate analysis of variance (MANOVA) was used.Of note, sTNFR2 was the in ammatory cytokine that exhibited the greatest change following in iximab versus placebo and was thus used as the primary immune endpoint as well as a proxy for TNF signaling in the statistical analyses (see Supplemental Materials).For longitudinal analyses, difference score distributions of sTNFR2 appeared parametric, and raw values were used for computation of difference scores unless otherwise noted.To assess change in anhedonia symptoms or effort discounting (k), an ANCOVA was used with drug condition as a xed factor and baseline values along with demographic variables as covariates.For choice behavior on the EBDM task, a repeated measures ANOVAs was used in lieu of an ANCOVA to model the interactions between time, drug condition and individual effort levels.Greenhouse-Geisser corrections are reported in cases where the sphericity assumption was violated.All statistical analyses included sex and age as covariates unless otherwise speci ed.All statistical analyses are two-tailed unless pre-registered as a one-tailed test or otherwise noted.

Neuroimaging Preprocessing and Data Quality Evaluation
For all neuroimaging preprocessing and rst-level GLM analysis, we used SPM12 (Wellcome Department of Imaging Neuroscience, Institute of Neurology, London, UK).SPM12 preprocessing included realignment estimation and implementation, co-registration to the individual's high resolution structural scan, normalization to MNI space, and spatial smoothing using a Gaussian kernel (6mm FWHM).To control for motion and other artifacts, data were visually inspected, and 6 realignment parameters were included as covariates for all GLM analysis.Visual inspection revealed 2 participants showing evidence of signi cant motion (> 3mm) that were subsequently examined as potential high-in uence data points.Additionally, the GLM contrasts used in univariate and multivariate analyses were evaluated for multivariate outliers using the Mahalanobis distance.The impact of subjects or sessions agged as potential outliers using this criterion were evaluated in a series of sensitivity analyses that included participants with poor motion and/or multivariate outliers as covariates and they were not found to alter the signi cance of any reported neuroimaging associations (see Supplemental Materials for Sensitivity Analyses).
Neuroimaging Analysis -First Level General Linear Models (GLM): For all rst-level GLMs, the canonical HRF was used, and event durations were modeled based on the duration of each cue for each trial with SPM default orthogonalization.Based on our pre-registered analyses and our prior work 11,39 , we examined change (14 day vs. baseline scans) across 7 parametric modulator contrasts: effort level at Cue 1, reward magnitude at Cue 1, predicted subjective value at Cue 1 (SV predicted ), choice di culty at Cue 2, subjective value of the chosen option (SV chosen ) at Cue 2, a subjective value prediction error (SVPE) and choice "shifts" at Cue 2 (Fig. 1B).
The parametric modulator contrasts for Cue 1 were de ned as follows: Contrast 1: Effort magnitude.The amount of effort required for a given trial (20%, 50%, etc.).Contrast 2: Reward magnitude.The amount of reward offered for a given trial ($1.00~$5.00).
Contrast 3: SV predicted .As in our prior work 11,39 , SV predicted was de ned using a sliding window analysis of previously-experienced subjective values of the same trial type.Therefore, the SV predicted for a trial that began by presenting 20% effort at Cue 1 would be calculated as the running average of the most recent SV values for trials that included 20% effort.
The parametric modulator contrasts for Cue 2 were de ned as follows: Contrast 4: SV chosen .The SV chosen was calculated as the subjective value of the option Contrast 5: SVPE.The subjective value prediction error (SVPE) regressor was estimated by calculating the absolute value of the difference between the SV chosen , and the SV predicted .
Contrast 6: Choice Di culty.A parametric modulator calculated as the difference between the subjective value of the effortful and non-effortful options.Choices for which this difference was small indicated greater choice di culty (because of similar values for both options).Contrast 7: Choice Shift.A parametric modulator coding a "1" if the choice on the prior trial was the same as the choice made on the current trial, and "0" if not.
A full description of results from all a priori contrasts is included in the Supplemental Materials.
In addition to the parametric modulators described above, we estimated single-trial models for the purpose of multivariate, beta-series and brain signature analyses.For these analyses, a trial speci c betaweight was estimated for the Cue1 onset, the Cue2 onset and the decision-phase onset of each trial.Analyses using single trial models focused on the Cue1 timepoint.

Neuroimaging Analysis -Second Level Univariate GLMs
For comparison of Day 14 and baseline scans, the two task runs from each timepoint were concatenated into a single rst-level GLM.A second-level contrast of [-1 -1 1 1] was then used to compare parametric modulators during the two baseline timepoint runs to the two 14 day timepoint runs.To examine the effects of in iximab, change in acceptance of 100% effort trials, change in effort discounting (k) and change in sTNFR2 levels, each of these variables were separately entered into second-level GLM that also included sex and age as covariates.For whole-brain analysis, correction for multiple comparisons was obtained using cluster-correction as implemented in SPM12, with an uncorrected height threshold of p < 0.001.For ROI mean betaweights from all voxels in each ROI was averaged and analyzed.
Neuroimaging Analysis -Univariate ROI Analysis: Our analysis plan identi ed ve a priori regions of interest: the dmPFC, bilateral insula, nucleus accumbens (NAcc) and ventromedial prefrontal cortex (vmPFC).Masks for these ROIs were drawn from a prior functional parcellation of medial prefrontal cortex 46 , the Glasser atlas 47 and the Harvard-Oxford atlas 48 .For ROI analyses, the rst eigenvariate of mean betaweights from all voxels within each mask was extracted for each participant and used in subsequent analysis.

Neuroimaging Analysis -Multivariate ROI Analysis
As an exploratory analysis, change in striatal sub-regions for the SV predicted and SV chosen was used to classify drug condition assignment.A cross-validated 3-fold classi cation with partial least squares regression was employed to estimate the area under the receiver operating characteristic (AUROC) for drug classi cation based on activity in bilateral ventral striatum, putamen and caudate as de ned by the Pauli basal-ganglia atlas 49 .Only neuroimaging data was used (i.e., age and sex covariates were not included).The k-fold procedure was estimated 10 times for each region, with the mean AUROC across each fold and iteration used to estimate predictive performance for each region.For statistical inference, we estimated the mean squared error (mse) of classi cation for each ROI and compared it to a null distribution of 5,000 mse values generated by a randomly permuted neuroimaging data.

Neuroimaging Analysis -Beta Series Correlation (BSC)
To understand how in iximab-induced changes in dmPFC responses during EBDM may drive networklevel changes within corticostriatal circuits, we examined task-based functional connectivity (beta-series correlation; BSC).To assess changes in BSC, we rst isolated the time series of beta-weights at Cue 1 for each ROI for each participant.We focused on Cue 1 for this analysis as it would detect changes prior to decision-outcome, and thereby could reveal network-level changes related to processing of partial information that would not be confounded by differences in the proportion of effortful options accepted between the two drug conditions.We then estimated the Pearson correlation between dmPFC and target striatal regions (VS, putamen, caudate).Resulting correlations were Fisher transformed to create a difference score in the BSC between each pair of nodes (ΔBSC).These ΔBSC scores were used as dependent variables in Ordinary Least Squares (OLS) regression analyses and bootstrapped-mediation analyses as described above with drug condition or ΔsTNFR2 used as predictor variables.
Neuroimaging Analysis -Reward Signature Comparison The goal of this analysis was to evaluate the similarity between change in neural activity during EBDM (14 day -baseline) and a pre-trained 'reward signature' developed to classify monetary wins during losses during gambling and monetary incentive delay tasks 50 .We rst calculated the change in cosine similarity between each participant's 14daybaseline mean contrast image and a neural signature.Regression was used to assess the relationship between the resulting cosine similarity and drug condition, ΔsTNFR2, Δk, and Δ100% Effort as well as self-report measures of anhedonia.

RESULTS
Effects of in iximab on behavioral and clinical measures of motivational anhedonia.
Using a 2 (treatment group) × 2 (time) × 4 (Effort level) repeated measures ANOVA, we observed a 3-way interaction (F (3,34) = 3.702, p = 0.033, η = .098)such that individuals receiving in iximab accepted signi cantly more effortful options at the higher effort levels (80% and 100%) relative to the placebo group (Fig. 2A).Interestingly, these group differences at the 14-day time-point were driven by both an increase in acceptance of effortful options at 100% effort among patients receiving in iximab (Cohen's d = .30)as well as a decrease among the placebo group (Cohen's d = .33).There were no consistent baseline differences between groups.
To further characterize the impact of in iximab on effort-based choice, we modeled effort discounting by tting a well-validated computational model of subjective value (see Supplemental Materials for modeling details).Consistent with the effects on effort-based choice behavior, we observed a signi cant (one-tailed) effect of drug condition on change in free parameter k that quanti es the amount that a reward is discounted by the effort required to obtain it (i.e.effort discounting; F (1,37) = 3.950, p = 0.028, η = .107)(Fig. 2B).Individuals receiving in iximab showed a signi cant decrease in k parameter values, indicating a reduction in effort discounting.

Associations between change in effort-based decisionmaking and a marker of TNF activity
In two separate linear regression models, we found that greater decreases in ΔsTNFR2 were associated with larger increases in acceptance rates for 100% effort trials (b=-0.461,p = .010)and larger reductions in effort discounting k parameter (b = 0.34, p = 0.030) (pre-registered as one-tailed) (Fig. 2C-D).Moreover, bootstrapped mediation analysis con rmed that ΔsTNFR2 partially mediated the relationship between drug condition (in iximab or placebo) and Δ100%Effort (95% CI: .007,.249)and Δk parameter (95% CI:-1.15,− .011)(see Supplemental Materials for full mediation results) Effects of in iximab on univariate imaging activity during effort-based decision-making and associations with behavioral and immunologic change.
We identi ed seven contrasts of interest for the EBDM task (see Supplemental Materials).For each ROI, we tested each contrast for effects of drug (in iximab or placebo), and change in circulating sTNFR2 (ΔsTNFR2).We did not observe a main effect of drug in our ROI analyses, but did observe an effect of sTNFR2 on encoding of choice di culty in dmPFC such that greater engagement of dmPFC during di cult choices was associated with a larger reduction in sTNFR2 (b=-.47,p = .013).Moreover, ΔsTNFR2 mediated the association between drug condition and choice-di culty beta weights (95% CI: .076,1.94).Full results from other planned contrasts are presented in Supplemental Materials.
We further observed effects of ΔsTNFR2 on dmPFC responses to subjective value of the chosen option SV chosen (see Methods).As with the choice di culty contrast, a bootstrapped mediation analysis found that ΔsTNFR2 mediated the relationship between drug condition and the change in dmPFC SV chosen encoding (95% CI: − .34,− .03).Additionally, change in SVC encoding mediated the relationship between ΔsTNFR2 and change in effort discounting (Δk) (95% CI: .000;.0048).Finally, using a whole-brain, clustercorrected analysis, we observed that dmPFC activity was signi cantly associated with ΔsTNFR2, change in effort discounting (Δk), and change in 100% effort choices (Δ100% Effort).Speci cally, change in the dmPFC was more strongly engaged by low subjective value trials in patients with larger sTNFR2 reductions (Fig. 3A), lower effort discounting (Fig. 3B) and more 100% effort choices (Fig. 3C).
All maps showed substantial overlap with each other and with our a priori ROI de nition for dmPFC (Fig. 3D & 3E).Though not predicted a priori, other areas showing whole-brain association included left motor cortex and aspects of medial frontal gyrus (see Supplemental Materials for results).
Effects of in iximab on multivariate encoding of subjective value.
To further probe for the effects of in iximab on subjective value information, an exploratory multivariate approach was used in three striatal ROIs: ventral striatum, putamen and caudate.Multivariate patterns were estimated using 3-fold cross-validated partial least squares (PLS) regression model on each subject's difference map for SV predicted value at Cue1 or SV chosen at Cue2.AUC classi cation between PLS predictions of drug condition and true condition assignment (in iximab and placebo) were then computed (see Supplemental Materials for further details).We observed strong evidence for classi cation in ventral striatum (AUC = .75,permutation test p = 0.007), with moderate classi cation in putamen (AUC = .68,permutation test p = 0.044) and no evidence for classi cation in the caudate (AUC = .55,permutation test p = 0.65) (Fig. 4A).Striatal regions did not show evidence for classi cation for activity during SV chosen .Taken together, these data suggest that changes in the patterns of activity within the ventral striatum and putamen systematically differed between treatment groups, thereby providing evidence of drug-related effects in these areas despite the absence of clear univariate changes.

Effects of in iximab on corticostriatal network processing
A limitation of the prior analyses is the focus on individual brain areas despite our knowledge that these areas function as an integrated circuit.To examine effects of in iximab on corticostriatal networks, we used an exploratory beta-series correlation (BSC) analysis 51 to identify how the task-dependent functional connectivity within the network was altered by drug condition and ΔsTNFR2.Given our observed effects on dmPFC at time of choice, this analysis focused on functional connectivity with dmPFC during Cue 1, as this might be the time when partial information was in uencing dmPFC connectivity to guide choice behavior.We tested dmPFC beta-series connectivity to ventral striatum (VS), putamen and anterior insula.As above, we found that decreasing sTNFR2 was associated with a signi cant decrease in beta-series connectivity (BSC) between dmPFC and bilateral putamen (Cue 1 reward: b = .53,p = .005;Cue 1 effort: b = .65,p = .0002)(Fig. 4B).Additionally, the effect of sTNFR2 mediated the association between drug condition and changing connectivity (Fig. 4C).There was no association with VS, and the change in BSC was signi cantly stronger for dmPFC-putamen than for dmPFC-VS (Steiger's Z = 3.46, p < 0.001).Taken together, these data suggest that altering sTNFR2 impacted network-level processing of reward and effort information even before full choice information was available.Effects of in iximab on "neural signatures" for reward.
In a nal set of exploratory analyses, we sought to assess whether a recently validated whole-brain predictive model (aka "neural signature") that was trained to decode gains versus losses during a monetary reward task 50 detected treatment-related changes in fMRI signals during EBDM.First, we determined sought to establish that a signature trained to differentiate win and loss feedback would be sensitive to high and low subjective value trials in our task.Using a previously collected sample of healthy controls, we observed a large difference in cosine similarity between high and low subjective value trials for the reward signature (Cohen's d = 1.05, t 46 =-5.024, p = 8e-6), with no difference for a control signature developed to classify negatively valanced-stimuli 52 (Fig. 5A; See Supplemental Materials).Next, we tested whether change in the pattern of whole-brain activity during Cue 2 of the EBDM task was associated with behavioral or symptom change.Consistent with our interpretation of brain-related changes implying greater motivation for reward, we found that Δ100%Effort choices made by participants were associated with a positive shift in cosine similarity with the reward signature such that individuals accepting more 100% effort options at day 14 relative to baseline also exhibited change towards away from a "loss-like" and more towards a "win-like" state (b = .32,p = 0.051).This effect was also associated with a greater reduction in the anhedonia subscale of the IDS-SR (b=-.44,p = .007)(Fig. 5B-C).

DISCUSSION
Consistent with expectations from our prior theoretical work 21,53 , we found that relative to placebo, in iximab was associated with an increase willingness to expend effort for rewards.This change was associated with enhanced subjective value encoding for effort-based choices within the dmPFC and the ventral striatum, two key structures that underlie human effort-based decisions [11][12][13][54][55][56] . Of note however, the observed changes in motivation measured by our behavioral task in in iximab-versus placebo-treated patients were associated with only nominally greater improvement in self-reported motivational anhedonia symptoms in in iximab-treated subjects.Together, these results support the hypothesis that anti-in ammatory strategies may be useful for targeting motivational behavior in depressed patients with evidence of chronic in ammation.
All participants randomized in our study met criteria for current MD and showed evidence of chronic in ammation (CRP > 3.0 mg/L) 57 and no evidence of infection, autoimmune and in ammatory disorders or other overt causes of in ammation.Substantial prior work has suggested that in ammation may be causally associated with MD 21,58,59 , but only for a subset of individuals.This suggests that some MD patients exhibit a vulnerability to chronic, low-grade in ammatory signaling that makes them more likely to develop motivational impairments, which can in turn exacerbate or maintain a depressive episode.The overarching goal of this study was to shed light on the neural circuitry that may underlie this vulnerability.
To that end, the rst hypothesis of our study was whether administration of an anti-in ammatory agent like in iximab would lead to increased motivation for rewards.This hypothesis was supported (Figure .2) with an overall reduction in the discounting of reward by effort, with the strongest effects of in iximab occurring at the highest levels of effort.This suggests that in iximab enhanced individuals' willingness to exert more strenuous levels of effort for rewards, rather than simply increasing the frequency with which they committed to moderate levels of effort.This effect was driven by changes in both groups; while patients treated with in iximab increased their acceptance (Cohen's d = .30),there was also a general decrease in willingness to expend effort among placebo-treated patients (Cohen's d = − .33); a pattern similar to that observed for placebo in a clinical trial using a monetary reward task and a kappa opioid antagonist 60 .
Regarding neural markers of this behavioral change, neuroimaging analyses highlighted a role for the dmPFC, such that individuals treated with in iximab exhibited stronger univariate responses with increasing choice di culty and subjective value.The dmPFC has long been known to play a critical role in EBDM, and the observed effects of ΔsTNFR2 showed substantial overlap with our prior work using this fMRI paradigm, demonstrating that dmPFC activity increases as the subjective value of the chosen option becomes less attractive (Fig. 3B). 11,39 he neural computations that underlie this region remain a subject of debate 17,61,62 .Indeed, some prior work has suggested that the association between dmPFC activity and subjective value may re ect the need to shift strategies 63 or greater di culty in adjudicating between the two options 17 , both of which frequently correlate with subjective value.In either case, our results suggest greater engagement in these valuation processes among individuals receiving in iximab.
The localization to dmPFC is further notable given prior evidence suggests that in ammatory stimuli directly alter measures of metabolism and activation in this region 24,27 , and that post-mortem samples from depressed patients show evidence of enhanced immune signaling [64][65][66] .Taken together, our results highlight dmPFC as a key structure mediating the impact of in ammation on motivated decision-making, and as a potential site of intervention for future treatments targeting the in ammatory subtype.
A possible caveat to this interpretation is that in iximab acted primarily on other brain areas, which led to a behavioral change re ected by changed dmPFC activity.In an effort to address this concern, we included two additional exploratory analyses: multivariate encoding of predictive subjective value (SV predicted ) at Cue 1 and changes in functional connectivity at Cue 1.The advantage of examining Cue 1 activity is that it allows for interrogation of brain responses to relevant cost/bene t information that are independent of participant choices, which cannot occur until full information is presented at Cue 2. Both these analyses found clear evidence supporting the effect of in iximab and change in TNF signaling on corticostriatal circuit function in a choice-independent manner.
A nal exploratory analysis was used to evaluate the effect of in iximab on whole-brain patterns of activity in our task using a previously-validated predictive brain model (aka "neural signature") that has been shown to discriminate between monetary gains from losses 50 .We found that patients accepting more 100% effortful options and endorsing lower anhedonia exhibited a shift away from a "loss-like" pattern of activity, possibly suggesting that effort costs were perceived to be less aversive.

Limitations
While the current study had many strengths, including a double-blind, randomized, placebo-controlled design, careful inclusion/exclusion criteria, and state-of-the art neuroimaging methods, several weaknesses warrant comment.First, due to the COVID-19 pandemic and related recruitment challenges, the resulting sample was only powered to detect large and medium effects.A second limitation was use of a single-dose and a relatively brief treatment window, which may have contributed to the lack of a clear treatment group × time interaction for self-reported measures of motivational symptoms.Importantly, the lack of an interaction was largely due to improvement across both groups, rather than an absence of improvement for those receiving in iximab.Moreover, the effect sizes were larger for in iximab for most measures.Given that we observed a clear treatment group × effort discounting interaction, it is possible that patients were beginning to experience more energy and motivation to pursue rewards, but that the placebo effect was still too strong to observe differentiation when patients were asked to report on their symptom experiences.

Conclusion
A single dose of in iximab resulted in greater willingness to exert effort for rewards, an effect that was re ected in effects on TNF signaling and corticostriatal circuitry, particularly in the dmPFC.These results further support the use of anti-in ammatory treatment strategies for targeting a possible "immunesubtype" of motivational impairments in patients with MD and other psychiatric disorders.

Declarations Figures
Effects of in iximab on effort-based decision-making.A. Group × Time × Effort level interaction in EBDM task performance.Relative to the placebo group, patients receiving in iximab accepted more effortful options at the two highest effort levels at 14 days.B. Group × Time interaction on effort discounting.Relative to the placebo group, patients receiving in iximab showed a reduction in effort discounting (indicated by the parameter k) at 14 days.C. A reduction in sTNFR2 at 14 days was associated with increased willingness to accept effortful options at the highest effort level (100%).D. A reduction in sTNFR2 at 14 days was associated with a reduction in effort discounting (k).These associations mediated the relationship between drug condition and effort performance.E. Changes in the MDFI Reduced Motivation Subscale for each group, with lower scores indicating greater improvement.F. Changes in the IDS-SR anhedonia subscale for each group, with lower scores indicating greater improvement.Abbreviations: MDFI -multidimensional fatigue inventory; IDS-SR -Inventory for Depressive Symptomatology Self Report; sTNFR2 -soluble TNF receptor 2; EBDM -Effort based decision-making.
Participants 131 were assessed for eligibility, and 42 experiencing a current MD episode as determined by Structured Clinical Interview for DSM-5 38 and a CRP > 3mg/L were randomized (see Figure S1 for consort diagram).