In February, 2023, an Ambio LIfe Science (“Ambio”) ibogaine treatment facility in Tijuana, Mexico received a 41-year-old male special forces veteran (“PA”) who was diagnosed with relapsing-remitting MS (RRMS), post-traumatic stress disorder (PTSD), major depressive disorder (MDD), and a traumatic brain injury (TBI) sustained ten years prior. He attended our program once previously, in September, 2022, at which time he suffered from vertigo believed to result from high alcohol intake. After his first treatment, vertigo symptoms resolved in the absence of alcohol consumption, only to return 2 months later, leading to his RRMS diagnosis.

At intake, PA’s symptoms included progressive mobility impairment, coordination difficulties, some bladder control issues, bradypsychia, short-term memory deficits, and strong vertigo sensations every five to ten minutes. Emotional stress from relationship dissatisfaction was an additional factor influencing his overall health. His primary care physician prescribed dimethyl fumarate and Vitamin D to manage MS symptoms one month prior, but progressive neurological decline continued. Magnetic resonance imaging (MRI) revealed a large lesion in the right posterior parietal lobe, affecting cortical and subcortical white matter near the intraparietal sulcus ( Figure 1 ).

Also in February 2023, Ambio received a 44-year-old female (“PB”) who was diagnosed with secondary progressive MS (SPMS) in 2018, as well as complex PTSD (CPTSD) from childhood trauma and a recent divorce following 17 years of marriage. Upon arrival, she exhibited severe muscle spasticity, hypotrophic extremities, difficulties with bladder and bowel control, and strong but infrequent attacks of vertigo. Although she could transfer herself using wall-mounted arm supports, she required a wheelchair for most activities. As of 4 months prior to intake, she was diagnosed “exercise intolerant”, precluding physiotherapy for more than 5 to 10 minute increments.

PB became confined to her mobility system after the installation of an intrathecal baclofen pump in 2019, which reduced the painful muscle spasticity that had previously allowed her to walk. She had taken glatiramer, high doses of Vitamin D, and also consumed cannabis on a regular basis to control muscle spasticity since 2018. Despite short-term success with ketamine therapy to try to control chronic pain, she terminated treatment infusions one year prior.

DWI data underwent initial preprocessing steps. Brain extraction was performed on the trace image using BET2 (Brain Extraction Tool) of the FMRIB software library [FSL; (28)]. The resulting brain mask was applied to the calculated ADC image to create a brain-extracted ADC.

Cortical surface reconstruction and cortical thickness measurements were performed using the recon-all-clinical pipeline from FreeSurfer, optimized for heterogeneous clinical MRI scans (29). The pipeline incorporates machine learning-based segmentation methods, such as SynthSeg (30, 31) and SynthSR (32), to enhance robustness and reliability. SynthSeg allows for segmentation of brain MRI scans of any contrast and resolution, making it suitable for comparative analysis of heterogeneous clinical brain scans. SynthSR uses super-resolution to convert anatomical scans of varying contrasts into T1-weighted images suitable for cortical segmentation. The Desikan-Killiany atlas (33) was employed for cortical parcellation.

Lesion segmentation and quantification was performed manually with the 3D Slicer software suite (34). The lesion was outlined on the T1 anatomical MR image at each slice (35) to create a detailed lesion segmentation mask as the measurement region of interest (ROI). The lesion ROI is then subsequently projected to its intra-subject DWI space for ADC quantification.

To align the anatomical (T1 or T2) images with the DWI space, a multi-stage registration process was implemented using ANTs (Advanced Normalization Tools) (36). The process involved the following steps: 1. Intra-subject longitudinal Rigid registration in anatomical space, from intrasubject time 2 to time 1. Diffeomorphic registration (SyN algorithm) from the DWI space to the anatomical space for each time point (37, 38). The ROIs for both cortical and lesion segmentations were then projected from anatomical to DWI space for ADC quantification.

A Gaussian Mixture Model was used to perform clustering analysis on inter-hemispheric cortical thickness percentage changes to identify patterns of regional alterations and possible changes in neurocircuitry.

One day after treatment, PA subjectively noted a resolution of MS symptoms, including motor and bladder issues. 2 months post-treatment, MSQLI fatigue subscores dropped 92%. Bladder control issues completely resolved. The physical components summary score increased 24%, and mental components summary score improved by 42% ( Table 1 ). Despite previous challenges walking because of an inability to coordinate foot movement, PA later reported participation in a 200 mile ultramarathon. One year after this second treatment episode, he still had not experienced any remission of vertigo.

Despite acute increases in muscle spasticity under ibogaine’s effects, PB reported reductions in spasticity post-treatment. HAI scores improved from 8 (“restricted to wheelchair”) at baseline, to 7 (“walking limited to several steps with bilateral support”) immediately upon discharge, a change sustained at month 2 follow-up. PB increased the duration of physiotherapy from 10 minutes to 1 hour, and reported continued gradual improvement 2 years later. MSQLI scores improved 73% for chronic pain, and 29% for fatigue. Bladder control score improved from 5 to 3, and bowel control score improved from 6 to 2. Physical components summary score improved 38%, while mental components summary improved 22% ( Table 1 ).

Post-treatment analysis revealed a 71% reduction in lesion volume (from 1,659.8 mm³ to 480.5 mm³) and a 35.6% decrease in mean ADC (from 2,756.3 × 10⁻⁶ mm²/s to 1,774.2 × 10⁻⁶ mm²/s), suggesting reduced extracellular diffusion ( Figure 1 ).

Regions associated with emotional and cognitive processing exhibited cortical thinning, including the left rostral and caudal anterior cingulate cortices (ACC; 6.94% and 5.23%, respectively), left frontal pole (4.66%), and right precuneus (4.01%). In contrast, cortical thickening was noted in the right cuneus (4.43%) and entorhinal cortex (3.67%), with asymmetric decreases in the left cuneus (3.38%) and entorhinal cortex (1.96%) ( Figure 2 ).

ADC changes varied across regions. Notable reductions were seen in the left lateral orbitofrontal cortex (4.96%) and right temporal pole (4.40%). Significant ADC increases included the right rostral middle frontal cortex (25.77%) and paracentral lobule (16.20%), and the left entorhinal cortex (11.13%) ( Figure 2 ).

Ventricular volume decreased bilaterally (~31%), with asymmetrical ADC changes (left: -25.02%; right: +4.92%). The left pallidum (48.3%) and amygdala (25.12%) increased in volume, while hippocampal changes were asymmetric (right: +20.49%; left: -11.08%). These changes suggest hemispheric compensatory responses.

Gaussian Mixture Model analysis of cortical thickness changes identified three clusters ( Figure 3 ). Regions associated with cortical thinning included areas involved in emotional processing, such as the ACC and insula. Regions showing symmetrical thickening or stability included areas related to memory and sensory integration, such as the middle temporal gyri and posterior cingulate cortex (PCC). Asymmetrical changes were observed in regions like the cuneus and entorhinal cortices, indicating differential hemispheric responses.

Cortical thinning was observed in several regions, including the right transverse temporal cortex (-9.17%) and left (-6.32%), and the right postcentral cortex (-4.10%) and left (-3.39%). Asymmetrical changes were notable in the entorhinal cortex, where the right decreased (-4.82%) and the left increased (+2.4%), and in the rostral ACC, where the left decreased (-4.82%) and the right increased (+6.01%). Conversely, cortical thickening was evident in regions such as the right isthmus cingulate (+5.52%) and left lingual cortex (+8.10%) ( Figure 2 ).

ADC changes varied across regions. Decreases included the left pars orbitalis (-5.72%), right pericalcarine cortex (-5.30%), and left lateral occipital cortex (-4.92%). In contrast, ADC increases were more prominent, such as in the left paracentral lobule (+22.10%), right paracentral lobule (+14.8%), left (+16.78%), and right PCC (+8.46%). Other notable increases included the right frontal pole (+14.34%), left entorhinal cortex (+15.17%), and caudal ACC (left: +13.27%, right: +10.73%) ( Figure 2 ).

Subcortical volume changes were also significant. The left and right lateral ventricle volumes decreased (-15.73% and -14.2%, respectively), while the right inferior lateral ventricle volume increased (+34.09%) as the left decreased (-5.74%). Asymmetrical changes were seen in the amygdala, with left volume decreasing (-9.51%) and right increasing (+20.19%), and in the accumbens area, with left volume increasing (+16.67%) and right decreasing (-16.47%). Increased volume was observed in the ventral diencephalon (right: +24.63%, left: +14.18%), pallidum (left: +13.96%, right: +12.02%), and putamen (left: +7.3%, right: +13.49%). Although caudate volume remained unchanged, ADC increased on both sides (left: +9.33%, right: +17.68%) ( Figure 2 ).

Clustering analysis identified distinct patterns of cortical thickness changes ( Figure 3 ). Regions with symmetrical thinning included clusters 1 and 4, encompassing areas associated with motor and executive functions, such as the paracentral lobules, postcentral gyrus, and transverse temporal cortex. These changes may reflect synaptic pruning and network optimization. Regions with symmetrical thickening, such as cluster 0, involved areas related to memory and executive functions, including the temporal pole and lateral orbitofrontal cortex. Asymmetrical left-side thickening and right-side thinning (cluster 2) were observed in memory and visual processing regions, including the entorhinal cortex, fusiform gyrus, and parahippocampal gyrus, possibly indicating neurogenesis and enhanced connectivity. Conversely, right-side thickening and left-side thinning (clusters 3 and 5) were prominent in regions associated with emotional regulation, including the anterior and isthmus cingulate cortices and medial orbital frontal cortex, suggesting differential hemispheric responses in emotional networks.