METHOD FOR EARLY DIAGNOSIS AND PERSONALIZED CHARACTERIZATION OF DRUG AND ALCOHOL ADDICTION FIELD OF THE INVENTION. The present invention relates to methods, markers, kits and models for early diagnosis and personalized characterization of drug and alcohol addiction. BACKGROUND Substance addiction is a chronic and relapsing neuropsychiatric disorder with significant health, social, and economic impacts. It is traditionally diagnosed through clinical evaluation by a psychiatrist, neurologist, or physician, and available treatment is limited. Typically, the evaluation relies on diagnostic criteria from the Diagnostic and Statistical Manual of Mental Disorders 5 (DSM5) to render diagnosis such as substance use disorder (SUD) or alcohol use disorder (AUD). SUD and AUD are characterized by compulsive drug/alcohol-seeking despite the harmful consequences. SUD and AUD affect physical and mental health and the ability to function, and are of the most common mental disorders globally. In the current absence of reliable physiological biomarkers for addictions, diagnosis of SUD/AUD are primarily determined by clinical psychiatric assessment. As disclosed herein, several growth factors in the brains of those affected by drug and alcohol use disorder have been implicated. (see Figure 1). However, objective means for diagnosing addictions such as SUD and AUD with physiological or biological markers has remained important but elusive. The present disclosure aims to provide a method for defining objective, measurable biomarkers to allow for an improved, more objective, and precise diagnostics tool, in addition to enabling more personalized treatment options. SUMMARY Aspects of the present invention provide a method for the early prediction of, diagnosis and/or personalized characterization of drug and alcohol addiction, to biomarkers for early prediction or diagnosis of drug or alcohol addition, in vitro tests using said biomarkers and to a kit for performing said method. The method may employ a computational model based on the expression data of select growth factors and, optionally, brain alterations reflected by scans, such as MRI including DTI scans. The method, predictive algorithm, and kit provide an objective means for measuring growth factors associated with chemical addiction. In a preferred embodiment, the computational model correlates the expression profiles of growth factors GDNF, BDNF, and FGF2 found in serum, blood or saliva with a diagnosis and personalized characterization of drug and alcohol addiction to provide a method for predicting the risk of developing addiction and strategies for treatment. According to a first aspect of the present invention, there is provided a biomarker composition for the prediction, diagnosis and/or personal characterization of a subject to risk of chemical addition, the composition comprising antibodies to at least to one growth factor selected from BDNF, GDNF and FGF2. Preferably, the biomarker composition includes antibodies to at least two of the growth factors, more preferably all three growth factors. Preferably, the antibodies include labels to allow at least one, preferably all of, qualification, classification and quantification of binding levels of the growth factors to the antibodies. A second aspect of the present invention provides a method for the in vitro prediction, diagnosis and/or personal characterization of a subject to a chemical addiction, the method comprising obtaining a sample of at least one of blood and/or saliva from a test subject; measuring the amount of growth factors in the sample; and comparing the amount of growth factors in the sample against a reference profile to determine risk of the subject to chemical addiction. Preferably, the growth factors are selected from at least one, preferably all, of BDNF, GDNF, and FGF2. Preferably, the method is carried out across time intervals to assess changing levels of the growth factors. More preferably, the method is combined with software to analyze the data obtained and provide an output on the potential risk of the subject to addiction. The method may optionally include assessing microstructural changes or alterations in a brain of the subject, for example by imaging techniques capable of depicting microstructural brain measurements, such as MRI (DTI) scans. Preferably, the biomarker data is correlated with these microstructural brain measurements. In a preferred embodiment of the method, the peripheral measurements of growth factors are correlated to a predicted brain change from alcohol or drug exposure via a predictive algorithm to make a prediction, diagnosis and/or personalized characterization of drug or alcohol addiction. A third aspect of the present invention provides a computer system configured to perform the method of the second aspect of the present invention, thereby providing computer-aided prediction and/or diagnosis of a subject to alcohol or substance addiction. The method may be implemented by a processor for making a diagnosis and personalized characterization of drug or alcohol addiction, comprising: creating a model dataset from an early-stage and advanced-stage model groups of subjects with a history of drug or alcohol exposure, said model dataset comprising information about brain levels of growth factors and optionally microstructural brain measurements (such as DTI-MRI), at earlier stages of alcohol or drug exposure and peripheral levels of growth factors at later stages of alcohol or drug exposure; training a predictive algorithm with the model data set by correlating brain levels of growth factors with peripheral levels of growth factors to create a predictive algorithm configured to correlate peripheral measurements of growth factors to predicted brain changes from alcohol or drug exposure to make a diagnosis and personalized characterization of drug or alcohol addiction; obtaining test data from at least one test subject, said test data comprising peripheral measurements of growth factor levels; and inputting test data into the predictive algorithm and outputting from the predictive algorithm a prediction of behavioral and brain changes from alcohol or drug exposure to make a diagnosis and personalized characterization of drug or alcohol addiction. Another aspect of the invention provides a processor and a non-transitory storage medium storing computer readable instructions, the processor configured to execute a predictive algorithm configured to correlate peripheral measurements of growth factors and clinical/behavioral measurements to predicted brain changes from alcohol or drug exposure to make a diagnosis and personalized characterization of drug or alcohol addiction. Yet a further aspect of the present invention provides an in vitro diagnostic kit for performing the method according to the second aspect of the present invention, the kit comprising: a means for obtaining a sample of blood or saliva from a test subject; a means for taking measurements of growth factor levels in the sample to obtain test data; and comparing the amount of growth factors in the sample against a reference profile to determine risk of the subject to chemical addiction The kit may include a probe. Preferably, the kit includes a processor for implementing a predictive algorithm configured to correlate peripheral measurements of growth factors to predicted brain changes from alcohol or drug exposure to make a diagnosis and personalized characterization of drug or alcohol addiction. For example, the processor may be incorporated into the probe or may be provided in a separate unit in communication with the probe. The present disclosure relates to a computational model based on the expression data of select growth factors and optionally microstructural brain changes reflected by imaging techniques capable of depicting microstructural brain measurements, such as MRI or DTI scans. According to the present disclosure, this computational model provides a method for predicting the risk of developing addiction at the early stages, and the optimal treatments at later stages of addiction. The findings according to the present disclosure can lead to a better understanding of the mechanisms that control SUD and AUD and the development of a new, objective approach for diagnosis and treatment of chemical addiction. BRIEF DESCRIPTION OF THE DRAWINGS Fig. 1. illustrates the GO and STOP pathways of addiction (adapted from Ron and Messing 2014, Barak and Ron 2016); Figures 2A to 2D demonstrates amphetamine induced increased psychomotor sensitization over 20 days. Mice spent 30 min in an open field arena and their spontaneous locomotor activity was recorded. Then, mice were injected with 2.5 mg/kg or 5 mg/kg amphetamine or vehicle, and returned to the apparatus for an additional hour. Fig. 2A shows number of rotations completed by mice for each session on the first day of injection. Fig. 2B shows number of rotations completed by mice for each session on the 10 ^th day of injection. Fig.2C shows number of rotations completed by mice for each session on the 20 ^th day of injection. Fig.2D shows a comparison by day of the mean number of the rotations completed by mice after injection. Data are presented in 5-min blocks of mean ± SEM. Bars represent mean ± SEM (n=7-8 per group). **p<0.01, ****p<0.0001; Figures 3A to 3C illustrate the effect of amphetamine administration on Fgf2, Bdnf and Gdnf relative expression following amphetamine injections in mice, showing FGF2, BDNF, and GDNF mRNA expressions in the medial prefrontal cortex (mPFC; Fig. 3A), nucleus accumbens (NAc; Fig. 3B), ventral tegmental area (VTA; Fig.3C) of mice after amphetamine sensitization protocol. Mice were administered with 2.5 or 5 mg/kg of amphetamine, or vehicle 10 times every other day. A day after the last injection, brains were extracted and dissected. Fgf2, Bdnf, and Gdnf expressions were determined using qRT-PCR and normalized to Gapdh. Each brain region’s data are normalized to their own vehicle control. Data are presented in 5-min bins of mean ± SEM. *p < 0.05, **p < 0.01, ***p < 0.001. (n = 7 – 8 per group); and Figures 4A to 4C demonstrate that amphetamine administration has minor effects on the serum growth factors level. Detection of the levels of FGF2 (Fig.4A), BDNF (Fig.4B), and GDNF (Fig.4C) in 0.5 µl of the mice serum during 20 days of amphetamine sensitization protocol. Mice were administered with 2.5 or 5 mg/kg of amphetamine, or saline 10 times every other day. Blood was withdrawn for protein analysis on the 1 ^st , 9 ^th and 21 ^st day . FGF2, BDNF, and GDNF expressions were determined using Dot blot and normalized to ponceau S. Each growth factor data are normalized to their own vehicle control. Bars represent mean ± SEM.(n=7-8 per group). Figures 5A to 5C demonstrate a longitudinal alcohol study investigating the effect of long-term alcohol consumption on mRNA expression levels of Fgf2, Bdnf and Gdnf in the mesocorticolimbic system, namely, the medial prefrontal cortex (mPFC; Fig. 5A), nucleus accumbens (NAc; Fig. 5B), and ventral tegmental area (VTA; Fig.5C) of mice after 12 weeks of alcohol consumption. Mice received access to 20% alcohol three days a week intermittently for 12 weeks. A day after the last alcohol consumption, brains were extracted and dissected. Fgf2, Bdnf, and Gdnf expressions were determined using qRT-PCR and normalized to Gapdh and expressed as percentage of their expression in water-drinking control. Bars represent mean ± SEM. #p<0.075,*p<0.05, **p<0.01(n=21-24 per group); Figures 6A to 6C demonstrate that 12 weeks of alcohol has minor effects on the serum growth factors level. Detection of the levels of FGF2 (Fig. 6A), BDNF (Fig. 6B), and GDNF (Fig. 6C) in 0.5 µl of the mice serum during 12 weeks of alcohol consumption. Mice received access to 20% alcohol three days a week intermittently for 12 weeks. Blood was withdrawn at 4 time points during the experiment: baseline (before drinking onset), and following 3, 6 and 12 weeks of alcohol consumption. FGF2, BDNF, and GDNF expressions were determined using Dot blot and normalized to ponceau S. Each growth factor data are normalized to their own vehicle control. Columns represent mean ± SEM (n=21-24); and Figures 7A to 7F illustrate how long-term alcohol consumption alters brain DTI indices. Figures 7 A-C. show FA values for each treatment group, extracted from the mPFC (Fig. 7A), striatum (Fig. 7B) and brainstem (Fig. 7C). Figures 7 D-F show MD values for each treatment group, extracted from the mPFC (Fig. 7D), striatum (Fig. 7E) and brainstem (Fig. 7F). Line graphs represent means ± SEM, expressed as change from the baseline time point. *p<0.05, **p<0.01, ***p<0.001 , ****p<0.0001 (n= 21-24). MD=mean diffusivity; FA= Fractional anisotropy; mPFC=medial prefrontal cortex; DETAILED DESCRIPTION OF EMBODIMENTS The present disclosure identifies and characterizes a set of biomarkers for substance addiction, not only in the brain but also in the periphery (blood and/or saliva samples), to allow non-invasive early diagnosis, intervention, and tailored therapy. The present invention provides biomarkers for substance addiction in the serum, and/or saliva, and tests their interaction with brain alterations. Microstructural alterations in the mouse brain have also been characterized, including changes in the brain white matter, and in connectivity between brain regions using Diffusion Tensor Imaging (DTI-MRI). Computational tools are then used to generate a model analyzing the growth factor expression data and the brain alterations DTI data, enabling prediction of the risk of developing addiction, early diagnosis, prognosis, and personalized treatment. These findings in mice may then be used to evaluate the effects of alcohol and drug addiction on the expression of FGF2, GDNF, and BDNF in the serum and saliva of humans, and to implement the model generated in animals in humans. The present invention demonstrates that FGF2, GDNF, and BDNF in the brain, serum, and/or saliva of mice can be useful biomarkers of the clinical expression of SUD and AUD. Specific microstructural changes in the brain MRI of mice can be useful biomarkers of the clinical expression of SUD/AUD. Computational modeling using both MRI indices and peripheral growth factor measurements allow for clinical diagnosis and prognosis for SUD/AUD. Over the past 15 years, several growth factors have been identified as reactive to drug and alcohol exposure, and as regulators of drug and alcohol related behaviors that characterize addiction. Specifically, according to the present disclosure, the levels of brain-derived neurotrophic factor (BDNF), glial cell line-derived neurotrophic factor (GDNF), and fibroblast growth factor 2 (FGF2) have been found to fluctuate in mesolimbic and nigrostriatal brain regions related to addiction following drug or alcohol exposure. Manipulations related to these growth factors can reduce addiction-related behaviors. The aim of the present disclosure is to identify central and peripheral biomarkers for addiction, first by using animal models, and then in human addiction patients. These biomarkers will be the basis for development of a kit for peripheral test (blood/saliva), combined with a software/application for analyzing the data and providing output of the potential risk for addiction and optimal treatment. To this end, central and peripheral alterations in the three growth factors in the brain, blood and/or saliva of mice under protocols of alcohol and amphetamine exposure were identified. Longitudinal exposure protocols previously shown to induce brain neuroadaptations were used. In parallel, the brains of the mice were scanned via MRI (DTI) at several time points to detect microstructural, connectivity, and white matter alterations in the brain. A computational model was then generated to allow for the prediction of addiction- related behaviors using specific biomarkers. Following establishment of the biomarkers in animal models, the peripheral levels of the three growth factors in human addictions may be investigated to provide a computational model for generation of a biomarker kit for humans based on these computational tools for addiction. Target users include individuals with high familial, genetic, social, or other risk factors associated with the development of addiction, individuals with habits of substance consumption, and patients with alcohol or drug use disorders. The method and kit according to the present disclosure will allow non-invasive early diagnosis, early intervention, and tailored personalized therapy. Brain-derived neurotrophic factor (BDNF) is a growth factor that belongs to the neurotrophic family and regulates the growth and survival of neurons and synapses. BDNF mRNA and protein expression are increased by alcohol consumption, but are reduced following long-term consumption and in response to withdrawal. Manipulation of BDNF can regulate alcohol consumption. For example, shRNA-mediated downregulation of BDNF in the dorsal striatum reduces alcohol consumption. Moreover, BDNF levels are increased when alcohol-related memories that cause relapse are being changed to memories that prevent relapse suggesting, according to the present disclosure, that the growth factor plays an important protective role against alcohol relapse. Glial cell line-derived neurotrophic factor (GDNF) regulates the development and survival of various neurons, and is highly expressed in the central and peripheral nervous systems. GDNF activates the receptor tyrosine kinase RET in the presence of the coreceptor GDNF-family receptor α1 (GFRα1) and activates intracellular signaling cascades. GDNF mRNA and protein levels in the mesolimbic brain reward system are increased by short-term alcohol intake, but withdrawal from excessive alcohol intake leads to GDNF deficiency in mesolimbic brain regions. Furthermore, GDNF levels differ among low and high alcohol drinkers. According to the present disclosure, GDNF plays an essential role as a regulator of alcohol consumption and protects from development of excessive alcohol drinking, but heavy drinking eventually leads to a breakdown of its mechanism, resulting in the escalation of alcohol intake and addiction phenotypes. (Figure 2). Fibroblast Growth Factor 2 (FGF2) is expressed throughout the developing brain and plays an essential role in cell development, and migration, and in neuronal maintenance. FGF2 binds to a tyrosine kinase FGF receptor 1 (FGFR1). Excessive alcohol consumption leads to increased expression of FGF2 and FGFR1 in the dorsomedial striatum in rodents. Activation of FGF2-FGFR1 increases, whereas inhibition of its activity reduces alcohol intake. Therefore, FGF2 is a positive regulator of alcohol consumption. It was further found that mice with FGF2 deficiency consume less alcohol than wild-type mice. Alterations in the expression of FGF2 in the brain in exposure to stimulants was also observed. Additionally, preliminary data according to the present disclosure suggests that the retrieval of alcohol- related memories produces a unique fingerprint of microstructural modifications of brain plasticity in several mouse brain regions, including memory and reward-related structures (e.g., hippocampus, striatum and septum. (Figure 3). These alterations were detected using diffusion tensor imaging (DTI), a diffusion MRI framework which enables the detection of cellular rearrangements of white-grey matter modifications and as a result allows the localization of neuroplasticity over long and short time scales. DTI has been used to probe micro-structural modifications, and consequently, activity-induced plasticity of brain structures. Moreover, it was found that the DTI-reflected microstructural modifications correlated with the extent of alcohol seeking. Thus, the present disclosure suggests that it would be possible to identify microstructural features in the brain that will correlate with the risk of developing excessive, out-of-control alcohol drinking, serving as additional biomarkers. As noted above, it was found that alcohol drinking increases the expression of FGF2 and GDNF and their receptors in addiction-related brain regions. Mice with FGF2 deficiency consume less alcohol, suggesting that FGF2 may provide a biomarker for addiction. In contrast, mice with GDNF deficiency show higher relapse and are more sensitive to drug reward, and GDNF levels are different among low and high-alcohol drinking rodents, suggesting the relevance of this growth factor as a biomarker as well. Additionally, it was found that BDNF levels elevate due to alcohol consumption and reduced in response to withdrawal. Importantly, these growth factor or their derivates were shown to be detectible in the blood and/or saliva of humans. Thus, FGF2 can be detected in saliva, and can serve as a biomarker for anxiety. BDNF and GDNF can be detected in serum and can serve as biomarkers for depression. GDNF, BDNF and FGF2 methylation in the serum are suggested according to the present disclosure to correlate with alcohol addiction. Materials and Methods Animals C57BL/6 (20-25 g at the beginning of the experiments) were bred at Tel Aviv University and housed at 21– 23°C under a 12h light-dark cycle (lights on at 07:00, lights off at 19:00) with food and water available ad libitum. All experimental protocols conform to the guidelines of the Institutional Animal Care and Use Committee of Tel Aviv University and the guidelines of the NIH (A5010-01). All efforts were made to minimize the number of animals and their suffering. Reagents Fast SYBR Green Master Mix (catalog no.4385618), TRIzol reagent (catalog no.15596026) and RevertAid kit (catalog no. K1691) were supplied by Thermo-Fisher Scientific (Waltham, MA, USA). Protease inhibitor cocktail 1 (catalog no.539131) was purchased from Merck (by Mercury Ltd., Rosh Ha’ayin, Israel(. Chloroform supplied by Bio-Lab (Jerusalem, Israel). All DNA oligonucleotides (qRT-PCR primers (, isopropanol, glycogen and ponceau S were obtained from Sigma-Aldrich (Rehovot, Israel). pro- BDNF (5H8) (catalog no. sc-65514) and FGF2 antibody (C-2) (catalog no. sc- 74412) were purchased from Santa Cruz Biotechnology (Dallas, Texas, USA). Anti GDNF EPR 2714(N) (catalog no ab176564) was purchased from abcam(Cambridge,UK). Horse anti Mouse HRP (lumiminecense) (catalog no 7076) and Horse anti Rabbit (HRP for luminescence) WB (catalog no 7074) were purchased from Cell Signaling Technology(Danvers, MA, USA). Nitrocellulose membrane (0.45µm) was purchased from TAMAR(Mevaseret Zion, Israel). Ethanol absolute (catalog no. 052523), purchased from Bio-Lab Ltd (Jerusalem, Israel), was diluted to 20% alcohol (v/v) in tap water (for voluntary consumption). D- amphetamine was purchased from R&D Systems (Biotest Ltd, Kfar Saba, Israel). Intermittent access to 20% alcohol in two-bottle choice (IA2BC) After a week of habituation for individual housing, mice were trained to consume alcohol in the intermittent access to 20% alcohol in a 2-bottle choice (IA2BC) procedure, as previously described (Carnicella et al., 2014; Even-Chen et al., 2017; Even-Chen and Barak, 2019; Ziv et al., 2019; Even-Chen et al., 2022). Briefly, mice received three 24-h sessions of free access to 2-bottle choice per week (tap water and 20% alcohol v/v) on Sundays, Tuesdays, and Thursdays, with 24 or 48 h of alcohol-deprivation periods between the alcohol-drinking sessions. During the withdrawal periods, animals received only water. The position (left or right) of each solution was alternated between each session to control for side preference. Water and alcohol bottles were weighed before and after each alcohol-drinking session, and consumption levels were normalized to body weight. The training lasted 6-12 weeks before brain extraction. Amphetamine-induced Psychomotor sensitization protocl Amphetamine was dissolved in saline and injected to mice intraperitoneally (i.p.), in a volume of 10 ml/kg every other day at a dose of 2.5 mg/kg or 5 mg/kg, for 20 days (a total of 10 injections). For the locomotor activity recordings, we conducted separate sessions due to the availability of eight apparatuses. In each session, eight mice at a time were placed each one in an open field arena(50cm X 50cm) for 30 minutes to monitor spontaneous activity. They were then injected with either amphetamine or vehicle and returned to their respective arenas for an additional 60 minutes of observation. Mouse performance (locomotor activity), assessed by the number of rotations, was analyzed using EthoVision 11 video-tracking system Serum preparation Blood was collected from the mice's submandibular vein (cheek punch), with about 30-50µl blood from each animal. Blood was left at room temperature for 20 min for clot formation. Then samples were centrifuged at 7500 RPM for 15 minutes, resulted serum was separated, aliquoted, and frozen at -80 ⁰ C until use. Dot Blot assay For the dot blot analysis, serum samples were diluted in Tris-buffered saline (TBS), with total volumes kept to 5 µl, and spotted onto a nitrocellulose membrane. After allowing the membrane to dry for 5-10 minutes, the membrane was stained directly with Ponceau S solution for 2 minutes for general protein quantification. To remove the Ponceau S solution, the membrane was briefly washed with TBST and blotting was performed using a 5% milk solution in TBST. This was followed by the application of a primary antibody diluted in 5% milk-TBS solution, which was left to incubate for 1 hour. After this period, the membrane was washed three times with TBST. HRP secondary antibody was then applied for 1 hour, followed by another series of three washes with TBST. The ECL Supersignal was applied for 5 minutes, followed by exposure times of 3, 4, and 5 minutes. The general proteins and the growth factors levels were generated using the measured signal intensity by ImageJ software (NIH). Quantitative Reverse Transcriptase Polymerase Chain Reaction (qRT-PCR) The qRT-PCR procedure was conducted as we previously described (Even-Chen et al., 2017; Zipori et al., 2017; Even-Chen and Barak, 2019; Ziv et al., 2019; Goltseker et al., 2021; Goltseker et al., 2022; Goltseker et al., 2023). Following brain dissection, tissue samples were immediately snap-frozen in liquid nitrogen and stored at −80°C until use. Frozen tissues were mechanically homogenized in TRIzol reagent and total RNA was isolated from each sample according to the manufacturer's instructions. mRNA was reverse transcribed to cDNA using the Reverse Transcription System and RevertAid kit. Plates (96 wells) were prepared for SYBR Green cDNA analysis using Fast SYBR Master Mix. Thermal cycling was initiated with incubation at 95°C for 20s, followed by 40 cycles of PCR with the following conditions: Heating at 95°C for 3s and then 30s at 60°C. Samples were analyzed in duplicate with a Real-Time PCR System (StepOnePlus, Applied Biosystems, Foster City, CA, USA), and quantified using the ΔΔCt method against an internal control gene, Gapdh. Change in the mRNA expression of the experimental groups was calculated as a percentage of the control group. Specific primers for the detection of mRNA expression were ordered from Thermo Fisher Scientific and diluted to 10 mM in ultra-pure water according to the manufacturer’s instructions (See the detailed sequence in Table 1). Magnetic resonance imaging (MRI) acquisition MRI scans were conducted in the Strauss Computational Neuroimaging Center at Tel Aviv University, by a Bruker Biospec 7T/30 Scanner equipped with a 660mT/m gradient unit, using a cross coil configuration of 86mm transmitive Volume coil and mouse quadrature coil as a receiver. The MRI scan protocol included structural T2 weighted (T2w) images that were acquired with the rapid acquisition with relaxation enhancement sequence (RARE) and DTI acquisition with a Diffusion-Weighted Spin-Echo Echo-Planar- Imaging pulse sequence (DW-SE-EPI). T2w acquisition was done with the following parameters: TR=2500 ms; effective TE: 31.2 ms, RARE factor 8 with 8 repetitions. 20 coronal slices, 0.6 mm thick (no gaps), with an in-plane resolution of 0.0875 mm2 covering the entire brain, and lasts 4:00 min. For DTI we have used TR/TE=3000/20 ms, Δ/δ=10/2.5 ms, 2 EPI segments, 30 gradient directions with b-value at 1000 s/mm2, and three B0 images. 28 axial slices, 0.6 mm thick (no gaps), In-plane resolution was 0.175 mm2. The DTI acquisition took ~10:00 min. During the MRI sessions, anesthesia was induced and maintained using isoflurane (1.5%) in pure oxygen. Isoflurane was obtained from Piramal Critical Care (Bethlehem, PA, United States). A heating system was used to maintain the animal's body temperature and their respiration was monitored and maintained at 30–50 breaths/min using a pneumatic balloon positioned against the animal's chest. Total MRI protocol acquisition took ~40 min. DTI analysis and image post-processing In this study, the DTI dataset was first corrected for head movement and eddy current distortion using the ExploreDTI platform (Leemans et al., 2009) within Matlab (MathWorks). To perform the EC/EPI correction, the dataset underwent non-linear tensor estimation with a correction to the structural T2w image. For the data quality check, we first reviewed the DTI data by visually inspecting the slice images, followed by an inspection of the motion correction parameters. Following that, the corrected DTI data was transformed into atlas space using nonlinear registration. This allowed for the extraction of the Fractional Anisotropy (FA) and Mean Diffusivity (MD) values for each mouse brain. FA represents the normalized standard deviation of diffusivities in each voxel, providing sensitive information about the presence and integrity of white matter fibers(Pierpaoli et al., 1996; Assaf and Pasternak, 2008). MD represents the average diffusivity within each voxel, reflecting tissue density and water diffusion characteristics (Alexander et al., 2007; Assaf and Pasternak, 2008). Subsequently, regional average FA and MD values were extracted for each scan. The selection of brain regions aimed to encompass key anatomical and functional compartments related directly or indirectly to the reward system, including the nucleus accumbens, prefrontal cortex, hippocampus, and others (Koob, 2013). Experimental design and statistical analysis Sex distributed approximately equally across the experiments and was initially analyzed as a factor; however, all analyzes did not yield a main effect of sex or any interaction with other factors (p's > 0.05). Therefore, data were collapsed across this factor. Amphetamine sensitization experiment In this experiment, we tested the effects of amphetamine administration in psychomotor sensitization protocols on Fgf2, Bdnf and Gdnf expression in the serum and brain regions of the mesolimbic system after 20 days. In addition, we measured the effect of different doses of amphetamine on locomotor activity. Three month old mice were injected with amphetamine (2.5 or 5 mg/kg) or vehicle every other day over 20 days (10 injections in total). We measured the mice's locomotor activity at three time points: at the first, 5 ^th and last injection days. (timeline, Figure 4). At each locomotor activity measurement, mice were placed in an open field arena for 30 min of spontaneous activity monitoring. Then, they were injected with amphetamine or vehicle, and returned to the arena for additional 60 min. Blood was withdrawn for protein analysis on the 1 ^st , 9 ^th, and 21 ^st days. Brain tissues were extracted for mRNA analysis after the last blood withdrawal (21 ^st day). The number of rotations was analyzed by mixed model two-way ANOVA with Time (5-min blocks) as within-subjects factor and amphetamine dose (2.5, 5 mg/kg or vehicle) as a between-subject factor. LSD post hoc analysis followed significant effects. The protein levels in the serum were determined using dot blot and normalized to ponceau S. The expression of each growth factor is expressed as percentage of its own vehicle control. Protein levels data were analyzed by two-way mixed model ANOVA, with a Duration of treatment as a within-subjects factor and amphetamine Dose as a between-subjects factor. ANOVA was followed by LSD post hoc tests. The mRNA expressions of Fgf2, Bdnf and Gdnf were normalized to Gapdh expression. mRNA levels in each brain region are expressed as percentage of the vehicle-treated control group. mRNA data were analyzed by one-way ANOVA, with a between-subjects factor of Treatment. ANOVA was followed by Dunnett post hoc tests. Finally, to investigate the association between the brain mRNA expression, protein levels in the serum and the amphetamine induced sensitization, we conducted Pearson linear correlation tests and used multiple linear regression (MLR) to predict amphetamine sensitization using variables. Longitudinal alcohol experiment In this experiment, we investigated the impact of long-term voluntary alcohol consumption on FGF2, BDNF, and GDNF expression levels in both plasma and brain, as well as its effects on the physical microstructural properties of brain regions. To calculate alcohol drinking and preference, we recorded alcohol and water consumption, along with body weight. A control group consumed only water. Following five days of habituation to individual housing, mice underwent baseline MRI (DTI) scans on PND 50. They were then trained to consume alcohol according to the IA2BC procedure. Subsequent scans were conducted on PND 71, PND 92, and PND 134, which corresponded to 3, 6, and 12 weeks of alcohol consumption, respectively (see Figure 1 for timeline). At the end of the experiment (12 weeks), we extracted the brains to assess the mRNA expression levels of these growth factors in mesocorticolimbic brain regions, specifically the medial Prefrontal Cortex (mPFC), Nucleus Accumbens (NAc), and Ventral Tegmental Area (VTA). After completing the pre-processing steps for the DTI analysis, we assessed changes in the FA and MD in the mPFC, striatum, and brainstem (containing the midbrain), between the scans taken before and after 3, 6, and 12 weeks of alcohol consumption. These indices were calculated for each subject in the respective groups. The protein expression levels of FGF2, BDNF, and GDNF in the serum were determined using dot blot and normalized to ponceau S. Each growth factor's data was normalized to its own vehicle control. The DTI indices (FA and MD) and protein levels were analyzed using a two-way mixed model ANOVA with a between-subject factor of drinking Solution (Water, Alcohol) and a within-subject factor of drinking duration (Baseline, 3, 6, and 12 weeks). Significant effects were analyzed by Fisher’s LSD post hoc test. The mRNA expressions of Fgf2, Bdnf, and Gdnf after 12-week alcohol drinking in the were normalized to Gapdh expression, and are expressed as percentage of the corresponding vehicle-treated control group. mRNA data was analyzed by unpaired t test. Finally, to understand the connection between the brain modifications observed with DTI, the mRNA, and protein expression levels with the behavioral data (the alcohol drinking and preference), we conducted Pearson linear correlation tests and used multiple linear regression (MLR) to predict alcohol drinking using variables. Experiment 1: Analysis of growth factor blood and brain expression following amphetamine-induced psychomotor sensitization To identify biomarkers for amphetamine addiction-related neuroadaptations, an amphetamine-induced locomotor sensitization protocol was used in which mice were injected with amphetamine (2.5 or 5 mg/kg, or saline) every other day over 10 and 20 days (10 injections). In the 10-day experiment mice locomotor activity was measured at two time points: the first and last injection days, and in the 20-day experiment, a measurement was added after the 5 ^th injection. (Time-lines shown at the top of Figures 2, 3 and 4). At each locomotor activity measure, mice were placed in an open field arena for 30 min of spontaneous activity monitoring. Mice were then injected with amphetamine or saline, and immediately placed back in the arena for additional 60 min. In the 10-day experiment, blood was withdrawn for protein analysis on the 1 ^st and 11 ^th day, in which brain tissues were extracted for mRNA analysis. In the 20-day experiment blood was withdrawn for protein analysis on the 1 ^st , 9 ^th and 21 ^st day, in which brain tissues were extracted for mRNA analysis. The amphetamine treated groups showed higher locomotor activity after the injection on the first day (see Figures. 2A, and 2C), with the 5 mg/kg amphetamine treated group were rotating significantly more than the 2.5 mg/kg amphetamine group. Mixed model two-way ANOVA: 10-day experiment: main effects of Dose (F[2,26]=32.51, p<0.0001) and Time blocks (F[17,442]=30.93, p<0.0001) and a Dose X Time blocks interaction (F[34,442]=25.43 , p<0.0001); 20-day experiment: main effects of Dose (F[2,20]=31.63 , p<0.0001) and Time blocks (F[17,340]=8.763, p<0.0001) and a Dose X Time blocks interaction (F[34,340]=10.79 , p<0.0001). Post hoc after injection: significant difference between both amphetamine groups and the vehicle group (p<0.05), and between the 2.5 and 5 mg/kg amphetamine group (p<0.05). In the locomotor activity tests following the amphetamine sensitization protocol (10 ^th and 20 ^th days), both the 2.5 mg/kg and the 5 mg/kg amphetamine groups showed a higher number of rotations and distance traveled after the injection compared to the vehicle, with no difference between the 2.5 mg/kg and the 5 mg/kg amphetamine groups (Figures. 2B). Mixed model two-way ANOVA: 10-day experiment, main effects for Dose (F[2,24]=34.19 , p<0.0001) and Time blocks (F[17,408]=37.52, p<0.0001), and a significant Dose x Time blocks interaction (F[34,408]=15.54 , p<0.0001); 20-day experiment, main effects for Dose in the 10 ^th day (F[2,20]=27.96, p<0.0001) and 20 ^th day (F[2,20]=24.10, p<0.0001) and of Time blocks in the 10 ^th day (F[17,340]=34.88, p<0.0001) and the 20 ^th day (F[17,340]=42.97, p<0.0001), as well as a significant Dose x Time blocks interaction in the 10 ^th day (F[34,340]=13.14 , p<0.0001) and in the 20 ^th day (F[34,340]=13.60, p<0.0001). Post hoc comparisons, 2.5 mg/kg or 5 mg/kg amphetamine vs. vehicle (p<0.05). Together these results indicate that the group treated with 2.5 mg/kg amphetamine underwent psychomotor sensitization. In the 5 mg/kg-treated group a ceiling effect might have been observed. Experiment 1A: Study Into the Effect of Amphetamine exposure in a sensitization protocol on mRNA expression levels of Fgf2, Bdnf and Gdnf in the mesolimbic system. The expression levels of mRNA of Fgf2, Bdnf and Gdnf in brain regions of the mesocorticolimbic system were investigated by dissection 24 h following the last day of an amphetamine treatment protocol that caused psychomotor sensitization, as described above. As shown in Figures 3A to 3C it was found that the mRNA expression levels of Fgf2 in the NAc and VTA were reduced in the 5 mg/kg group. One-way ANOVA: Fgf2 in the NAc (F[2,19]=3.916, p=0.0377, and post hoc 5 mg/kg vs. vehicle, p=0.0338, Fgf2 in VTA (F[2,19]=4.883, p=0.0194, and post hoc 5 mg/kg vs. vehicle, marginally significant - p=0.0516). Additional group differences did not reach statistical significance (p>0.05). Repeated amphetamine injections were found to alter the expression of the growth factors in the mesolimbic system, so investigations were carried out to examine whether amphetamine-induced psychomotor sensitization correlates with the levels of these growth factors. Experiment 1B: Amphetamine-induced psychomotor sensitization correlates with FGF2, BDNF and GDNF protein levels in the serum of mice. The levels of FGF2, BDNF and GDNF in 0.5 µl of the mice’s serum were measured before and during the amphetamine sensitization protocol, using dot blot (see Figures 4A to 4C). The levels of BDNF were found to increase with time, but no difference was detected between the groups. No other effects were found in either of the two experiments. Mixed model two-way ANOVA: FGF2, no main effects for Dose (F[2,20]=0.275, p=0.762) and Time (F[1,20]=0.868, p=0.363), and no significant Dose x Time interaction (F[2,20]=0.68, p=0.518); BDNF, no main effect for Dose (F[2,20]=0.732, p=0.493), significant main effect for Time (F[1,20]=4.608, p=0.044), and no significant Dose x Time interaction (F[2,20]=1.321, p=0.289) ); GDNF, no main effects for Dose (F[2,20]=0.099, p=0.905) and Time (F[1,20]=0.735, p=0.401), and no significant Dose x Time interaction (F[2,20]=2.007, p=0.161). The serum levels of FGF2 and GDNF were found to positively correlate with the sum of rotation (FGF2, r=0.709, p=0.074, n=7), (GDNF, r=0.794, p=0.033, n=7) at the 2.5 mg/kg dose after 10 days. While no correlations were found at the 2.5 mg/kg dose after 20 days of amphetamine sensitization training (|r|<0.173. p>0.05, Table 10), negative correlations were found between the serum levels of FGF2 and BDNF and the sum of rotation (r=-0.924, p=0.001, n=8; r=-0.708, p=0.049, n=8, respectively) at the 5 mg/kg dose after 20 days. Finally, the baseline (before training) serum levels of FGF2 and GDNF positively correlated with the sum of rotation (r=0.746, p=0.034, n=8; r=0.822,p=0.012, n=8, respectively) at the 5 mg/kg dose after 20 days. Experiment 1C: Combinations of blood growth factors levels can predict amphetamine- induced psychomotor sensitization indices The data was then analysed to find whether multiple growth factor data can predict the psychomotor sensitization induced by amphetamine better than each growth factor on its own, using multiple linear regression models. It was found that the baseline FGF2 and GDNF serum levels predicted rotation behavior following the first day of injection in the control group. The fitted regression model was: Rotation Behavior=162.389 +0.234*(baseline GDNF serum levels) – 0.792*(bassline FGF2 serum levels). As displayed in Table 1 below, this model could explain 80.5% of the variance in rotation behavior (R²=0.805, F[2, 5]=10.308, p=0.017). The baseline GDNF serum levels were found to significantly predicte rotation behavior (β=0.990, p=0.021). On the contrary, baseline FGF2 serum levels were found to negatively predict rotation behavior (β=-1.347, p=0.006) (Table 1 below). Table 1 - 1 ^st day rotation behavior prediction (control group): Unstandardized Standard Standardized Coefficients (B) Error Coefficients (Beta) t Stat P-value Intercept 162.389 25.383 6.398 0.001 For the 2.5 mg/kg treatment group, the overall regression model predicting rotation behavior on the first day based on baseline BDNF and FGF2 serum levels. The fitted regression model was: Rotation Behavior=353.798 – 2.042*(bassline FGF2 serum levels) +1.106*(bassline BDNF serum levels). The model was was statistically significant, explaining 93% of the rotation behavior variance (R²=0.932, F[2, 4]=27.611, p=.005). Baseline BDNF serum levels significantly predicted rotation behavior (β=0.707, p=0.014). Conversely, baseline FGF2 serum levels were found to negatively predict rotation behavior significantly (β=-1.247, p=0.002) (Table 2 below). Table 2 - 1 ^st day rotation behavior prediction (2.5 mg/kg group): Unstandardized Standard Standardized Coefficients (B) Error Coefficients (Beta) t Stat P-value Intercept 353.798 43.601 8.114 0.001 FGF2 Serum BL -2.042 0.276 -1.247 -7.402 0.002 For the 2.5 mg/kg treatment group, the overall regression model was used to predict rotation behavior on the 10 ^th day based on the baseline BDNF, FGF2, and GDNF serum levels. The fitted regression model was: Rotation Behavior=516.391 -0.644*(bassline GDNF serum levels)–1.16*(bassline FGF2 serum levels)+1.113*(baseline BDNF serum levels). The model was statistically significant and explained 97.8% of the variance in rotation behavior (R²=0.978, F[3, 4]=43.723, p=0.006). It was found that the baseline BDNF serum levels significantly positively predicted rotation behavior (β=0.555, p=0.032). However, the baseline GDNF (β=-0.938, p=0.004) and FGF2 (β=-1.247, p=0.007) serum levels were found to significantly negatively predict rotation behavior (Table 3). Table 3 – 10 ^th day rotation behavior prediction (2.5 mg/kg group): Unstandardized Standard Standardized Coefficients (B) Error Coefficients (Beta) t Stat P-value Intercept 516.391 38.647 13.362 0.001 For the 5 mg/kg treatment group, the overall regression model was used to predict rotation behavior on the 20 ^th day based on the FGF2 serum levels at the 10-day point, and GDNF serum levels at the 10 and 20-day points. The fitted regression model was: Rotation Behavior=1323.22+1.616*(GDNF serum levels on the 10 ^th day)–9.888*(FGF2 serum levels on the 10 ^th day) -1.2*(GDNF serum levels on the 20 ^th day). The model was statistically significant and explained 96% of the variance in rotation behavior (R²=0.960, F[3, 4]=31.823, p=0.003). GDNF serum levels at the 10-day point were found to significantly predict rotation behavior positively (β=1.697, p=0.002). However, the FGF2 serum levels at the 10-day point (β=-1.808, p=0.001) and GDNF serum levels at the 20-day point (β=-1.023, p=0.001) were found to significantly negatively predict rotation behavior (Table 4). Table 4 - 20 ^th day rotation behavior prediction (5 mg/kg group): Unstandardized Standardized Coefficients (B) Standard Error Coefficients (Beta) t Stat P-value Intercept 1323.220 118.351 11.181 0.000 These results suggest that growth factors at different time points can predict amphetamine locomotor sensitization. Experiment 1D: Non-invasive Prediction of Brain Growth Factors from Blood Measurements. A multiple linear regression analysis was also carried out to predict the mRNA expression of the three growth factors in the brain, using their blood protein levels. For the 2.5 mg/kg treatment group, the overall regression model was used to predict mRNA expression Gdnf levels in the mPFC based on the GDNF serum levels at baseline, the 10 ^th day, and the 20 ^th day. The fitted regression model was: Gdnf=709.675 -2.592*(bassline GDNF serum levels) -3.478*(GDNF serum levels on the 10 ^th day)-0.791*(GDNF serum levels on the 20 ^th day). The model, although explaining 98.5% of the variance , was not statistically significant (R²=0.985, F[3, 1]=21.990, p=0.155). Neither the baseline GDNF serum levels (β=-1.891, p=0.090), the GDNF serum levels at the 10 ^th day (β=-1.055, p=0.128), nor at the 20 ^th day (β=-1.350, p=0.082) significantly predicted Gdnf levels in the mPFC (Table 5 below). Table 5 – Gdnf mRNA expression levels prediction: Unstandardized Standard Standardized Coefficients (B) Error Coefficients (Beta) t Stat P-value Intercept 709.675 73.950 9.597 0.066 GDNF Serum BL -2.592 0.369 -1.891 -7.026 0.090 GDNF Serum 10d -3.478 0.711 -1.055 -4.890 0.128 GDNF Serum 20d -0.791 0.103 -1.350 -7.687 0.082 #p < 0.2 This result support the hypothesis that blood measurements can provide a non-invasive means of predicting brain growth factor levels. Thus, it is shown herein that the amphetamine sensitization protocol affects the expression levels of growth factors in the brain and periphery. Importantly, these changes in the growth factors expression further correlate with and predict this sensitization. An aspect of the present invention relates to the use of these growth factors as peripheral biomarkers for amphetamine addiction. Experiment 2: Investigation into Biomarkers for alcohol use disorder – a longitudinal study. To assess biomarkers for alcohol addiction, a longitudinal study was conducted in which mice consumed alcohol for 12 weeks (see Figures 11A-11C). Several central and peripheral measurements were taken in 4 time points during the experiment: baseline (before drinking onset), and following 3, 6 and 12 weeks of alcohol consumption. In each time point, blood and saliva were collected for growth factor detection, and the brains of mice were scanned via MRI (diffusion tensor imaging, DTI). At the end of the experiment (12 weeks of alcohol drinking), the brains of the mice were extracted to determine the growth factors mRNA levels. A control group drank water only. The DTI data, brain and blood samples were analyzed to determine microstructural alterations, connectivity and myeline adaptations, and central and peripheral changes in growth factor levels. Data was analyzed to form a computational model that can predict alcohol drinking phenotypes. Experiment 2A: Long-term alcohol consumption alters the mRNA expression levels of Fgf2 Bdnf and Gdnf in the mesocorticolimbic system. The mRNA expression of Fgf2, Bdnf and Gdnf in mesocorticolimbic brain regions system was analysed, dissected 24 h following the last day of the 12-week alcohol consumption period. As shown in Figure 5A, we found that in the mPFC, the expressions of both Fgf2 and Bdnf were lower in the alcohol group compared with their expressions in controls. Moreover, in the NAc (See Fig. 5B), Fgf2 and Bdnf expressions were lower, but Gdnf expression was higher compared to controls. In addition, in the VTA, the expression of both Fgf2 and Bdnf levels were lower in the alcohol group compared with their expression in controls (see Fig.5C). (mPFC: Fgf2 t(35)=2.071, p=0.046, Bdnf (t(38)=1.921, p=0.062;NAc: Fgf2 t(38)=3.286, p=0.002, Bdnf (t(35)=2.01, p=0.052, Gdnf (t(40)=2.265, p=0.029; VTA: Fgf2 t(38)=2.547, p=0.015, Bdnf t(35)=2.011, p=0.052) all other comparisons, p>0.075). Experiment 2B: Long-term alcohol consumption correlates with the mRNA expression levels of Bdnf and in the VTA. An investigation was carried out to identify potential biomarkers by testing whether there is a correlation between alcohol consumption and the expression of the growth factors in the mesocorticolimbic brain regions. We found a negative correlation between the mRNA expression levels of Bdnf in the VTA following 12 weeks of alcohol consumption and the alcohol intake levels at the same time (r=-0.582, p=0.009, n=19). Experiment 2C: Long-term alcohol consumption correlates with the serum levels of BDNF and GDNF protein in mice Next, the dot blot method was used to quantify the levels of FGF2, BDNF, and GDNF in the mice’s serum before and during alcohol consumption. Levels of BDNF and GDNF were found to change with time, but no difference was detected between the alcohol- and water-treated groups (Figure 6A-6C). Mixed model two-way ANOVA: FGF2, no main effects of Group (F[1,40]=0.023, p=0.879), and Time (F[3,120]=1.52, p=0.213), and no Group x Time interaction (F[3,120]=0.628, p=0.598); BDNF, a main effect for Time (F[3,120]=9.36, p<0.0001), but no main effect for Group (F[1,40]=0.027, p=0.869), and no Group x Time interaction (F[3,120=0.922, p=0.432); GDNF, a main effect for Time (F[3,117]=27.20, p<0.0001), but no main effect for Group (F[1,39]=0.806, p=0.375), and no Group x Time interaction (F[3,117=0.484, p=0.694). Importantly, a marginally-significant positive correlation was found between the GDNF serum levels and alcohol consumption after 3 weeks, but not at later time points (r=0.427, p=0.068, n=19). In addition, BDNF serum levels positively correlated with alcohol consumption after 12 weeks of alcohol intake (r=0.505, p=0.019, n=21). No correlations were found between serum levels of FGF2 and alcohol intake (p’s>0.05) (Tables 2-4). These results indicate that long-term alcohol consumption affects the expression levels of growth factors in mesocorticolimbic brain regions, and that the levels of the growth factor in the brain and blood correlate with the levels of alcohol intake. Experiment 2D: Long-term alcohol consumption alters DTI indices in the mPFC, striatum and brainstem. Microstructural changes in the brain were characterized using an in vivo neuroimaging protocol that employed a within-subjects longitudinal study design, in which the mice’s brains were scanned before, during and after the consumption of alcohol in the IA2BC procedure, at four time points (0, 3,6 and 12 weeks of alcohol intake). The Fractional Anisotropy (FA) and Mean Diffusivity (MD) values were extracted for each mouse scan. FA represents the normalized standard deviation of diffusivities in each voxel, providing sensitive information about the presence and integrity of white matter fibers, as is known in the art. MD represents the average diffusivity within each voxel, reflecting tissue density and water diffusion characteristics. The FA values in the striatum were found to change with time (see Figure 7B). This observation aligns with previous studies that found a development-related increase in FA. In addition, we found that the reduction from baseline in MD values were higher in the alcohol-drinking group compared with water-drinking controls, in the mPFC (Figure 7D) and the striatum (Figure 7E). These findings imply an increase in tissue density within these regions as a result of excessive and chronic alcohol intake. In the context of alcohol addiction, these results raise the possibility that the development of addiction- related behaviors may engage learning processes that alter the brain's microstructural attributes. It is reasonable to postulate that recurrent alcohol exposure and the associated reward experiences might initiate adaptive neural circuit changes, culminating in a more densely structured brain tissue. The MD values in the brainstem changed with time (Figure 7F) (FA -mPFC: no main effects of Drinking Solution (F[1,43]=0.288, p=0.594), and of Duration of drinking (F[2,86]=1.308, p=0.275), but a marginally significant Drinking Solution X Duration interaction (F[2,86]=2.717, p=0.072). Striatum: a main effect of Duration of drinking (F[2,86]=3.47, p=0.035), but no main effect of Drinking Solution (F[1,43]=0.835, p=0.366), and no Drinking Solution X Duration interaction (F[2,86]=0.291, p=0.748). Brainstem: no main effect of Drinking Solution (F[1,43]=0.946, p=0.336), and of Duration of drinking (F[2,86]=2.026, p=0.138) and no Drinking Solution X Duration interaction (F[2,86]=0864, p=0.425). MD - mPFC: main effects of Drinking Solution (F[1,43]=4.882, p=0.0325), and Duration of drinking (F[2,86]=9.171, p=0.0002) but no Drinking Solution X Duration interaction (F[2,86]=0.336, p=0.715). Striatum: main effects of Drinking Solution (F[1,43]=7.36, p=0.009), and Duration of drinking (F[2,86]=16.68, p<0.0001) but no Drinking Solution X Duration interaction (F[2,86]=0.603, p=0.549). Brainstem: a main effect of Duration of drinking (F[2,86]=3.736, p=0.028), but no main effects of Drinking Solution (F[1,43]=0.015, p=0.902), but no Drinking Solution X Duration interaction (F[2,86]=0.028, p=0.972)). These results indicate that long-term alcohol consumption alters cortical, striatal and brain stem levels of MD and FA. However, DTI indices can be affected by various other factors such as myelination, axon density, axon diameter and membrane permeability and therefore, it cannot be decisively concluded what these alterations in MD and FA reflect physiologically. Experiment 2E: Combinations of the growth factors levels and DTI measurements predict alcohol intake within each time point. Multiple linear regression models were then used to determine whether a combination of the growth factors and the DTI measurements can predict alcohol intake. At the three-week timepoint, it was found that the overall regression model, used to predict alcohol intake based on the three-week MD values of the mPFC and the serum protein levels of GDNF. The fitted regression model was: 3 Weeks Alcohol Intake=15.467 +0.019*(GDNF serum levels at three weeks)- 0.415*(MD values in the mPFC at three weeks). The model was statistically significant, explaining 32% of the variance in alcohol drinking (R²=0.322, F[2, 16]=3.804, p=0.045); (see Table 6 below). Table 6 - 3 weeks alcohol intake prediction: Unstandardized Standardized Coefficients (B) Standard Error Coefficients (Beta) t Stat P-value = At the 12-week time point, the predictive model, included serum levels of BDNF and FGF2. The fitted regression model was: 12 Weeks Alcohol Intake=7.561-0.132*(FGF2 serum levels at twelve weeks)+0.251*(BDNF serum levels at twelve weeks). The model was statistically significant and explained 59% of the variance in alcohol drinking (R²=0.593, F[2,18]=13.131, p<0.001). The serum levels of FGF2 and BDNF at the twelve-week time point significantly predicted alcohol intake (see Table 7 below) Table 7 – 12 weeks alcohol intake prediction: Unstandardized Standard Standardized Coefficients (B) Error Coefficients (Beta) t Stat P-value Intercept 7.561 4.369 1.731 0.101 12W -0.132 0.034 -0.793 -3.868 0.001 12W BDNF serum 0.251 0.049 1.044 5.095 <0.001 ** p < 0.001 Multiple linear regression was then used to predict alcohol drinking at the later phases of the training procedure (12 weeks) using predicting variables from the early stages of the procedure. First, measurements from the 6-week time point were combined with baseline FGF2 serum levels to predict alcohol intake at the 12-week point. The fitted regression model was: 12 Weeks Alcohol Intake=9.934+0.155*(BDNF serum levels at six weeks)-0.203*(FA values in the brainstem at six weeks)- 0.051*(baseline FGF2 serum levels). The overall regression model, was statistically significant, explaining 61% of the variance in alcohol drinking (R²=0.605, F[3,17]=8.675, p=0.001). All these predictors were found to significantly contribute to the combined prediction of alcohol intake at 12 weeks (see Table 8 below). Table 8– 12 weeks alcohol intake prediction: Unstandardized Standard Standardized Coefficients (B) Error Coefficients (Beta) t Stat P-value Intercept 9.934 4.240 2.343 0.032 In the second regression model, measurements of the FA values of the striatum and the MD values of the mPFC at the three-week were combined with baseline GDNF serum levels to predict alcohol intake. The fitted regression model was: 12 Weeks Alcohol Intake=26.413-0.113*(FA values in the striatum at three weeks)-0.030*(baseline GDNF serum levels)-0.402*(MD values in the mPFC at three weeks). The model was found to explain 35% of the variance in alcohol drinking with marginal significance (R²=0.353, F[3,17]=3.096, p=0.055). On their own, these predictors showed only marginal or no significant contribution to the overall prediction (see Table 9 below). Table 9 – 12 weeks alcohol intake prediction: Unstandardized Standard Standardized Coefficients (B) Error Coefficients (Beta) t Stat P-value Intercept 26.413 3.172 8.328 <0.001 Str-FA 3W -0.113 0.059 -0.382 -1.907 0.074 BL GDNF serum -0.030 0.016 -0.383 -1.958 0.067 mPFC-MD 3W -0.402 0.261 -0.306 -1.536 0.143 # p < 0.075 suggest the peripheral expression of growth factors as predictors of subsequent alcohol consumption at the long-term time point, suggesting that such a combination has strong potential to serve as biomarkers for AUD. The inventors have identified that long-term alcohol consumption and amphetamine sensitization protocol significantly affect the expression levels of Bdnf, Fgf2, and Gdnf in the mesolimbic brain system. These changes, as well as the expressions of the growth factors in the blood, correlate with the levels of alcohol consumption and the extent of amphetamine-induced psychomotor sensitization. In addition, using multiple linear regression models, it was found that combinations of the growth factors and DTI measurements at the early stages predicted alcohol intake after 12 weeks. Similarly, combinations of the growth factors' levels were effective in predicting amphetamine locomotor sensitization in certain groups and time points. Aspects of the invention relate to the combined detection of fluctuations in these growth factors levels in the blood, along with observed microstructural brain changes, serving as potential biomarkers for identifying the development and progression of excessive alcohol drinking and amphetamine locomotor sensitization in a mouse model. Currently, there are no reliable biomarkers of addiction, and diagnosis and treatments are determined solely by subjective psychiatric/neurological clinical assessment according to DSM-V criteria. The present disclosure identifies and characterizes a set of biomarkers for substance addiction, not only in the brain but also in the periphery, which allows for non-invasive early diagnosis, intervention, and tailored therapy. It is to be appreciated that while the experimental data relates to alcohol and amphetamine addition, it is not limited as such as may be used for the diagnosis and treatment of other addictions, such as other substances for example cocaine. According to certain embodiments, a kit is provided for early diagnosis of addiction (substance use disorder) in a non-invasive manner (using a blood and/or saliva sample), backed up by brain imaging as a complementary approach to refine and validate the diagnosis. According to the disclosure, computational tools have been used to analyze big data generated from behavioral measurements, molecular and physiological data from the periphery, and MRI data from brain scans. Using these computational tools, the optimal model has been developed to predict alcohol or drug addiction phenotypes at early stages, and to identify the affected brain system(s) using peripheral data, allowing more personalized therapy. According to some embodiments, the disclosure relates to a kit for providing a peripheral test (blood/saliva) for the determination of addiction phenotypes. There are currently no available reliable biomarkers for addiction. Therefore, the kit according to the present disclosure will provide the first such biomarker kit, allowing fast, non-invasive profiling of: 1. A patient's tendency to develop addiction (for early diagnosis of susceptible individuals); 2. The stage of addiction and its severity in an objective and reliable manner (for addiction patients); and 3. A subject’s personalized profile of addiction, allowing tailored treatment (e.g., potential substances of addiction, stage of the disorder, predicted relapse rates, specific brain mechanisms involved and most relevant treatments). According to other embodiments, software / applications / access to databases provide a means and method for analyzing the data and providing output of the potential risk for addiction, and optimal treatment. Target users can include individuals with high familial, genetic, social, or other risk factors for developing addiction; individuals with habits of substance consumption (hazardous alcohol drinkers or drug users); patients with alcohol or drug use disorder; and individuals with other maladaptive addictive habits (e.g., gambling, computer/mobile gaming). In certain cases, past tense has been used for purposes of reduction to practice where ongoing experimentation is being conducted. The demonstration of the relevance of these physiological biomarkers to addiction in mice also provides clear support for their potential role as biomarkers in human subjects