Altered patterns of fractional amplitude of low-frequency fluctuation and regional homogeneity in abstinent methamphetamine-dependent users

Methamphetamine (MA) could induce functional and structural brain alterations in dependent subjects. However, few studies have investigated resting-state activity in methamphetamine-dependent subjects (MADs). We aimed to investigate alterations of brain activity during resting-state in MADs using fractional amplitude of low-frequency fluctuation (fALFF) and regional homogeneity (ReHo). We analyzed fALFF and ReHo between MADs (n = 70) and healthy controls (HCs) (n = 84) and performed regression analysis using MA use variables. Compared to HCs, abstinent MADs showed increased fALFF and ReHo values in the bilateral striatum, decreased fALFF in the left inferior frontal gyrus, and decreased ReHo in the bilateral anterior cingulate cortex, sensorimotor cortex, and left precuneus. We also observed the fALFF values of bilateral striatum were positively correlated with the age of first MA use, and negatively correlated with the duration of MA use. The fALFF value of right striatum was also positively correlated with the duration of abstinence. The alterations of spontaneous cerebral activity in abstinent MADs may help us probe into the neurological pathophysiology underlying MA-related dysfunction and recovery. Since MADs with higher fALFF in the right striatum had shorter MA use and longer abstinence, the increased fALFF in the right striatum might implicate early recovery during abstinence.


Results
Demographic characteristic. Twelve participants (5 MADs and 7 HCs) were excluded for analysis due to excessive head motion. A total of 70 MADs and 84 HCs were included in the final analysis. The groups did not differ in gender, age, years of education, or head motion ( Table 1). The cigarette smoking per day (CPD) (p < 0.001) and body mass index (BMI) (p = 0.007) in MADs were significantly higher than those in HCs. The MADs had an abstinence period for 38.5 (26.47) days.
fALFF results between groups. Compared to HCs, MADs showed significantly increased fALFF in the left caudate extending to left putamen, right pallidum extending to right putamen, and right caudate (Fig. 1, Table 2). MADs showed significantly decreased fALFF in left inferior frontal gyrus (IFG) compared to HCs.
ReHo results between groups. MADs showed significant increased ReHo in the left caudate extending to left putamen, and the right putamen extending to right caudate, and decreased ReHo in the bilateral anterior Table 1. Demographic characteristics of subjects. *Median (interquartile range) and Mann-Whitney test are presented. # Mean (standard deviation) and two sample t-test are presented.

Discussion
This study is the first to compare the fALFF and ReHo differences between abstinent MADs and HCs in a large sample. Compared to HCs, abstinent MADs showed significantly increased fALFF and ReHo in the bilateral striatum (caudate and putamen), decreased fALFF in the left IFG and decreased ReHo in the bilateral ACC, precentral and postcentral gyrus, and left precuneus. Furthermore, fALFF values of the right striatum cluster and the left striatum cluster were negatively correlated with the natural log of the duration of MA use and positively correlated with age started using MA, with fALFF value of the right striatum cluster was positively correlated with the natural log duration of abstinence. A few of studies with a small sample size used a single method (ReHo [36][37][38] or ALFF 39 ) to explore resting state brain activity among MADs. Some studies had combined ALFF and ReHo among individuals with psychiatric Figure 1. fALFF analysis. Two-sample t-test results are presented, voxel-level p < 0.001, cluster level p < 0.05, two-tailed, voxel size > 9, corrected by GRF. The area in blue represents significantly decreased ALFF value; the area in yellow and red represents significantly increased fALFF value. Created with DPABI_V4.3_200401 (http:// rfmri. org/ dpabi).  [40][41][42] and other addiction disorders 43,44 , but not in MADs. As fALFF and ReHo could be complementary to each other, combining these two methods to explore resting state brain activity among a relatively large sample of MADs may reveal more information regarding to the pathophysiological framework in the brain during abstinence. In this study, increased fALFF and ReHo in the bilateral striatum for MADs may indicate higher amplitude of BOLD signals and increased spontaneous regional neural activity in these two brain regions. The altered fALFF and ReHo in the bilateral striatum not only revealed abnormal and non-synchronous spontaneous neural activity in these two regions but also highlights them as core brain regions that are altered in the resting brains of abstinent MADs. Only a few rs-fMRI studies have investigated the alternation of spontaneous brain activity and synchronization using fALFF or ReHo among individuals with addiction disorders. Active heroin users showed decreased ALFF in the right caudate, which was negatively correlated with the duration of heroin use 31 . A study compared heroin users with and without relapse found that relapsed heroin users had increased ReHo in the right caudate 45 . Among individuals with internet gaming disorder, ALFF values of the bilateral putamen were higher before cognitive based treatment (CBT), while ALFF of left putamen was lower after CBT compared to HCs 46 . Chronic active smokers showed significantly increased fALFF in the right caudate, relative to HCs 47,48 . Although our results are not quite consistent with these studies, which may be due to the heterogeneity of samples (i.e., different addiction, different state of patients at data acquisition, and different sample size), these results consistently indicate the bilateral striatum as key areas in addiction disorders, in line with previous studies. Furthermore, the values of fALFF and ReHo may change during the course of abstinence. Both addictions and obsessive-compulsive disorder (OCD) show abnormalities in striatum, which is related to impaired selfcontrol 49 . OCD patients exhibited lower ReHo in the bilateral caudate before treatment and increased ReHo in Figure 2. ReHo analysis. Two-sample t-test results are presented, voxel level p < 0.001, cluster level p < 0.05, two-tailed, voxel size > 29, corrected by GRF. Area in blue with significantly decreased ReHo value; area in yellow and red with significantly increased ReHo value. Created with DPABI_V4.3_200401 (http:// rfmri. org/ dpabi).  56,57 in abstinent MADs have been reported in previous studies. These findings revealed the important roles of striatum in the pathology of MA dependence. Since these results are observed during abstinence, some of the neurotoxic effects of MA appear to be persistent, while some studies reported partial recovery in MADs with protracted abstinence 54 , some even with rapid recovery 55 . Thus, the increased fALFF and ReHo values of bilateral striatum in the present study may indicate partial recovery of striatum function in abstinent MADs. Importantly, we found that the fALFF value of the right striatum was positively correlated with duration of abstinence, which may suggest that long-term abstinence is beneficial for the recovery of striatum function, such as self-control. MADs living in an environment where there is no access to MA do not have to make as many effort as those who have access to MA, thus more resources retained for better self-control 19 , which may be related to the increased fALFF value of the striatum. Moreover, the fALFF values of bilateral striatum were negatively correlated with the duration of MA use, which may further suggest that the increased fALFF values of bilateral striatum, especially the right striatum, are associated with abstinence, or even early recovery, rather than only MA impairments. Within this context, the present findings suggest that the fALFF and ReHo values of www.nature.com/scientificreports/ the bilateral striatum change over the course of abstinence. The fALFF values of bilateral striatum may decrease during active MA use and early stages of abstinence, possible due to MA acute effects, then increase with longer periods of abstinence due to partial recovery or a compensatory reaction. Thus, fALFF values of bilateral striatum, especially the right striatum, may be sensitive to the change of state between active use and abstinence. While the fALFF and ReHo values of bilateral striatum may not continually increase or even may decrease to levels the same as healthy controls at some timepoint, this is only a speculation as the current study did not have a long abstinent follow-up data to test this. Further work is needed to demonstrate these dynamic changes.
Adolescents are more vulnerable to drug use-related alterations than adults [58][59][60] . Animal studies showed a similar pattern whereby the younger the age of exposure to MA, the more vulnerable and long-lasting the impairment is 61,62 . These findings are largely consistent with our present results, in which the increased fALFF values of bilateral striatum were positively correlated with the age of first MA use. This indicates that the younger the age of first MA use, the less the fALFF recovery during abstinence.
In our study, MADs showed decreased fALFF in the left IFG. The IFG is thought to play an important role in response inhibition 63 . Both MADs and individuals with other addiction disorders exhibited lower activation in IFG during response inhibition-related task and decision-making task [64][65][66] , which correlated with impairment of inhibition control. Decreased grey matter volume of the left IFG was found in both MADs 67 and alcoholdependent patients compared to HCs 68 . Moreover, early abstinent MADs had reduced cerebral blood flow (CBF) of the IFG 69 . A previous study demonstrated both ReHo and ALFF are reliably correlated with regional CBF in most brain regions 70 . Therefore, decreased fALFF in the left IFG is suggestive of MA-related impairment.
The decreased ReHo in the bilateral ACC in the present study is consistent with previous studies among abstinent MADs 38 , alcohol use disorder subjects 71 , and betel quid dependence subjects 43 , compared to HCs. ACC is recognized as an important area in addiction. Previous studies of MADs have shown decreased glucose metabolism 72 , abnormal metabolite levels 59 , decreased cerebral blood flow 73 , and hypoactivation during decision-making tasks 74,75 , which are associated with deficits in attentional control, cognitive functions of behavior monitoring, risk-related processing, and self-control 14 . ACC is actively involved in cognitive control, emotional regulation, and self-control 76 . These results lead us to speculate that the functional changes in the ACC may underlie impaired cognitive function and self-control in MADs.
As a key functional region of the DMN 77,78 , the precuneus is involved in the awareness of the perception of environmental stimuli (exteroception) 79 . Exteroception contributes to hyper-sensitivity to self-relevant external cues associated with their drug use in addiction. In alignment with this theory, increased precuneus activation when exposed to drug cues is widely-reported in different addicted populations [80][81][82] , which elicit craving related to cue reactivity. Precuneus has been considered as a vulnerable region in addiction as it is a core region of the exteroception network 79 . For example, individuals with MA associated psychosis showed stronger ReHo in the right precuneus than individuals with schizophrenia 37 . The inconsistency with our present result is reasonable due to different samples. In addition, decreased glutamate + glutamine was observed in the precuneus in early abstinent MADs 83 . Cortical thickness trended smaller in precuneus in MADs 84 . Consequently, we speculate that the decreased ReHo in the left precuneus is related to dysregulation of the exteroception process due to MA use.
It is noteworthy that decreased ReHo was found in the bilateral precentral gyrus and postcentral gyrus, i.e., the sensorimotor cortex. The effects of drug exposure on the sensorimotor cortex have not been fully investigated yet. While smokers showed fALFF decreased left postcentral gyrus 85 , MADs showed decreased glucose metabolism 72 and lower grey matter volume in the precentral gyrus 86 . Although few studies have looked specifically at MA effects within the sensorimotor cortex, they have extensive connections. Previous studies report putamen-sensorimotor circuits play an important role in habitual responding 87 and relapse 88 . The abnormal ReHo values of bilateral precentral gyrus and postcentral gyrus may suggest that these regions play an important role in MA-related impairment.
There are several limitations in the neuroimaging study. First, this is a cross-sectional study, so we could not confirm the causal relationship between these alterations with MA consumption, or abstinence, or a combination of MA consumption and abstinence. Second, we focused on resting state and did not apply any cognitive tasks, so we could not show whether brain alterations were correlated with cognitive impairment. Third, the MADs and healthy controls were not matched exactly regarding BMI or CPD. During the abstinence, MADs could not access to MA, and may lead to withdrawal symptoms like anxiety; many would increase cigarette use, sleep a lot, decrease physical activity, and eat more snacks, which may increase body weight. Therefore, both BMI and CPD of MADs were higher than HCs. Although both CPD and BMI showed influence in ReHo and fALFF in several brain regions 47,85,[89][90][91] , these alterations could not explain all the results in our study. Furthermore, we included BMI and CPD as covariates to control these effects in the analysis. Finally, most of the subjects were male, a possible gender difference was unexplored. Therefore, future studies should recruit more female MADs, match subjects' CPD, and BMI, include cognitive function tests, and measure alterations during different stages (i.e., active MA use, early abstinence, and long-term abstinence).
Nevertheless, this first study showed abstinent MADs had abnormal resting-state function in cortical and subcortical regions in bilateral striatum, left IFG, bilateral ACC, precentral gyrus, postcentral gyrus, and left precuneus in a large sample. These brain alterations may relate to self-and cognitive control and exteroception. We also observed increased fALFF in striatum were correlated with duration of MA use, age of first MA use and duration of abstinence. We speculate that the decreased fALFF and ReHo may be the result of MA consumption and the increased fALFF and ReHo in the bilateral striatum may result from abstinence. These findings may help to elucidate the pathophysiology in MADs.  93 . MADs were excluded if they met criteria for other substance dependence (excluding nicotine dependence) at any time. Subjects were required to abstain from MA for at least 48 h before scanning. Drug-free HCs were recruited from the community through advertising. Participants were excluded if they (i) had any general medical condition or neurological disorders, including infectious, hepatic, or endocrine disease; (ii) had a history of severe head injury with skull fracture or loss of consciousness of more than 10 min; (iii) had any current or previous psychiatric disorder; (iv) had a family history of psychiatric disorder; (v) women during pregnancy or breast-feeding stage; (vi) had contraindications for MRI. Two licensed psychiatrists, at MD level, conducted all clinical interviews. Subjects were fully informed about the measurement and MRI scanning in the study. Written informed consent was given by all subjects. This study was approved by the Ethics Committee of the Second Xiangya Hospital, Central South University (No. S095, 2013), and was carried out in accordance with the Declaration of Helsinki. The demographic characteristics are shown in Table 1.
Magnetic resonance data acquisition. The images were acquired using standard sequences with a Sie- The functional images were collected using an echo-planar imaging sequence with the following parameters: number of slice = 32, repetition time = 2,000 ms, echo time = 30 ms, slice thickness = 4.00 mm, flip angle = 90°, matrix = 64 × 64, field of view = 220 × 220 mm 2 .

Data preprocessing. Functional and structural images were processed by Data Processing & Analysis of
Brain Imaging (DPABI) 94 using Data Processing Assistant for Resting-State fMRI (DPARSF) 95 . The first ten volumes were discarded to allow for signal stabilization and subjects adaptation. The remaining volumes were corrected for slice time differences, realigned to correct for small movements, and corrected for head motion. Subjects with head motion exceeding 2.0 mm in any dimension or 2° of any angular rotation were excluded from further analysis. Individual functional images were then coregistered to T1-weighted MR images, which were segmented and normalized to the standard structural MRI template in the Montreal Neurologic Institute space using nonlinear transformation. Spatially normalized images were then detrended to remove linear trends and remove nuisance signals, including white matter, cerebrospinal fluid signals, mean global signal, and 24 motion parameters.
Calculation of fALFF and ReHo. fALFF and ReHo values were calculated based on previous studies 32,33 using DPABI. For fALFF analysis, the detrended functional images were smoothed with a Gaussian kernel of 4 mm full-width at half-maximum (FWHM). Power spectrum were computed by transforming time series of each voxel to the frequency domain via Fast Fourier Transform. The average square root of the power spectrum at each voxel across 0.01-0.1 Hz was taken as the ALFF. fALFF was obtained as the division of ALFF by the whole frequency range observed in the signal.
ReHo calculation was performed on a voxel-by-voxel basis by calculating Kendall's coefficient of concordance (KCC) of the time series of a given voxel with its nearest 26 voxels 32 . Then the individual ReHo maps were smoothed with a Gaussian kernel of 4 mm FWHM.
A whole-brain mask was used to remove the nonbrain tissues. Prior to subsequent analyses, individual fALFF and ReHo maps were standardized into z-score maps by dividing the global mean fALFF and mean KCC within the whole-brain mask. Statistical analysis. Statistical analysis was performed in R 3.6.1 within Rstudio 96 . Differences between MADs and HCs in demographic variables, i.e., age, gender, duration of education, marriage, CPD, and head motion were tested using two-sample t-tests, Mann-Whitney tests, and Pearson's chi-squared tests using R, and a p < 0.05 was set as significant.
Analysis of fALFF and ReHo maps were performed with voxel-wise two-sample t-tests in DAPBI. We included age, CPD, BMI, head motion, and grey matter volume as covariates. Significant differences in the analysis were reported using the criteria of multiple comparisons with the Gaussian Random Field theory correction (GRF) (voxel-wise p < 0.001, cluster-wise p < 0.05, two-tailed) and with a minimum extent threshold of 30 voxels for ReHo, 10 voxels for fALFF.
Regions where the MADs showed significant differences over the HCs for ReHo or fALFF properties were determined as regions of interest (ROIs). ReHo/fALFF values of these regions were extracted, averaged, and regressed against MA use parameters (i.e., age started using MA, duration of MA use, duration of abstinence). MA use parameters that were not normally distributed were natural log-transformed for the regression analyses 51 . We corrected the model p value for each regression model by using Bonferroni correction 97 . Models with corrected p value less than 0.05 would be considered as there was a significant relationship between the value of Scientific Reports | (2021) 11:7705 | https://doi.org/10.1038/s41598-021-87185-z www.nature.com/scientificreports/ fALFF or ReHo and MA use parameters. Then Benjamini-Hochberg (BH) procedure was applied to control the false discovery rate (FDR) 98 to correct the p value of MA use parameters in the models with corrected model p value less than 0.05.