Whole-brain R1 predicts manganese exposure and biological effects in welders.

Area of Science:

• Neuroscience
• Radiology
• Occupational Health

Background:

• Manganese (Mn) is a neurotoxicant and a paramagnetic substance used as an MRI T1 contrast agent.
• Mn accumulation in the brain is linked to neurological disorders, including parkinsonian symptoms.

Purpose of the Study:

• To investigate the relationship between manganese exposure, brain manganese accumulation, and neurological indicators using machine learning.
• To determine if MRI R1 maps can identify individuals with higher manganese exposure and associated neurological changes.

Main Methods:

• Support vector machines (SVM) were trained using whole-brain R1 maps from 57 welders and 32 controls.
• Classification was based on air Mn concentration ([Mn]_Air), brain Mn accumulation (ExMn_Brain), motor dysfunction (UPDRS), and thalamic GABA concentration (GABA_Thal).
• SVM was also used to predict age to validate results against Mn exposure.

Main Results:

• R1 maps accurately predicted air Mn concentration above 0.20 mg/m³ (88.8% accuracy, 88.9% recall).
• R1 predicted lower thalamic GABA concentrations (≤ 2.6 mM) with 82% accuracy and 78.9% recall.
• R1 predicted age below 48 years (75-82% accuracy) but not above, suggesting Mn-specific signatures are linked to cumulative exposure.

Conclusions:

• Higher manganese air exposures and longer welding durations result in identifiable brain signatures detectable by SVM using R1 maps.
• Lower levels of manganese exposure (< 0.20 mg/m³ and < 18 years welding) do not produce discernible brain signatures.
• MRI R1 mapping combined with SVM is a promising tool for assessing occupational manganese exposure and its neurological impact.