Zoe Cohen

EECS Department, University of California, Berkeley

Technical Report No. UCB/EECS-2024-39

May 1, 2024

http://www2.eecs.berkeley.edu/Pubs/TechRpts/2024/EECS-2024-39.pdf

Iron is vital for proper functioning of the brain. Deficiency of iron can impair the formation of myelin sheaths, disrupting critical communication between neurons and causing disorder in movement, mood, attention, energy levels, and more. Although iron is critical for many neural processes, it is a delicate balance to maintain a healthy amount of iron. When unbound in the cell, ferrous iron can serve as a catalyst in the formation of reactive oxygen species which can cause DNA damage and cellular death in excess. Even in normal aging, iron accumulates in deep grey matter areas that are known to exhibit damage in neurodegenerative disorders like Alzheimer’s and Parkinson’s disease. We propose a mathematical model to describe the network transmission of iron throughout the brain over the lifespan, resulting in the accumulation seen in the deep grey nuclei.

The model is a linear Markov model with an input term, and the parameters are estimated using iron concentration estimates from Quantitative Susceptibility Mapping (QSM) Magnetic Resonance Imaging (MRI). We discuss the motivation behind using QSM for model training and how parameters were estimated. We also present an evaluation of the model performance and validation performed using a population collected separately from the training dataset. Finally, we share results for an application case of the model by comparing the model prediction when trained on healthy controls and subjects diagnosed with either Alzheimer’s or mild cognitive impairment. These results show that our model is consistent in predicting iron transport dynamics for separate healthy populations and also predicts altered dynamics when trained on pathological QSM data.

Our model is able to predict iron transport dynamics that accurately reproduce the spatial and temporal distribution of iron seen over the lifespan. This work serves to help unravel the mystery of iron accumulation in the brain, a process which has implications for the treatment and early diagnosis of neurodegenerative disorders.

Advisors: Chunlei Liu

---

BibTeX citation:

@phdthesis{Cohen:EECS-2024-39,
    Author= {Cohen, Zoe},
    Title= {Modeling System-level Iron Homeostasis in the Human Brain over the Lifespan},
    School= {EECS Department, University of California, Berkeley},
    Year= {2024},
    Month= {May},
    Url= {http://www2.eecs.berkeley.edu/Pubs/TechRpts/2024/EECS-2024-39.html},
    Number= {UCB/EECS-2024-39},
    Abstract= {Iron is vital for proper functioning of the brain. Deficiency of iron can impair the formation of myelin sheaths, disrupting critical communication between neurons and causing disorder in movement, mood, attention, energy levels, and more. Although iron is critical for many neural processes, it is a delicate balance to maintain a healthy amount of iron. When unbound in the cell, ferrous iron can serve as a catalyst in the formation of reactive oxygen species which can cause DNA damage and cellular death in excess. Even in normal aging, iron accumulates in deep grey matter areas that are known to exhibit damage in neurodegenerative disorders like Alzheimer’s and Parkinson’s disease. We propose a mathematical model to describe the network transmission of iron throughout the brain over the lifespan, resulting in the accumulation seen in the deep grey nuclei. 

The model is a linear Markov model with an input term, and the parameters are estimated using iron concentration estimates from Quantitative Susceptibility Mapping (QSM) Magnetic Resonance Imaging (MRI). We discuss the motivation behind using QSM for model training and how parameters were estimated. We also present an evaluation of the model performance and validation performed using a population collected separately from the training dataset. Finally, we share results for an application case of the model by comparing the model prediction when trained on healthy controls and subjects diagnosed with either Alzheimer’s or mild cognitive impairment. These results show that our model is consistent in predicting iron transport dynamics for separate healthy populations and also predicts altered dynamics when trained on pathological QSM data. 

Our model is able to predict iron transport dynamics that accurately reproduce the spatial and temporal distribution of iron seen over the lifespan. This work serves to help unravel the mystery of iron accumulation in the brain, a process which has implications for the treatment and early diagnosis of neurodegenerative disorders.},
}

EndNote citation:

%0 Thesis
%A Cohen, Zoe 
%T Modeling System-level Iron Homeostasis in the Human Brain over the Lifespan
%I EECS Department, University of California, Berkeley
%D 2024
%8 May 1
%@ UCB/EECS-2024-39
%U http://www2.eecs.berkeley.edu/Pubs/TechRpts/2024/EECS-2024-39.html
%F Cohen:EECS-2024-39