Neural Functional Asymmetry
Looking at associations between neural functional asymmetry and Alzheimer's Disease progression
At the Geriatric Psychiatry Neuroimaging (GPN) Lab, I have worked with my research partner Li Jinghang, and mentors Harold Aizenstein, M.D/PhD, Helmet Karim, PhD, and Akiko Mizuno, PhD, on characterizing the relationship between neural functional asymmetry and Alzheimer's Disease (AD). progression.
​
We generate MATLAB programs to quantify neural activation patterns, Python to organize and create files for participant data, and R to assess statistical significance of relationships between functional asymmetry and AD factors. We have one manuscript pending review for Ingenium, an undergraduate journal at the University of Pittsburgh, and a preprint publication on the way.
​
My clinical experience at the GPN Lab includes phone screenings for an anxiety research study. I have also helped establish, and now facilitate an "Intergenerational Conversations" course through the Virtual Senior Academy to help build intergenerational connections.
Objective
To find the statistical significance of associations between extent of dedifferentiation (i.e. symmetry and spread of activation) and AD related factors (i.e. APOE E4 status, global amyloid deposition, and global glucose metabolism)
Intro
-
Alzheimer’s disease (AD) is a process of progressive neurodegeneration induced by cytotoxic elements such as amyloid-β deposition (Aβ).
​
-
AD is associated with lower glucose metabolism and AD risk is associated with the APOE E4 allele
​
-
Greater pathology may be associated with compensatory increases in neural activation, such as increased spread of activation (i.e., greater neural resources) and greater bilateral activation.
​
-
This is known as dedifferentiation, defined as a loss of brain functional specificity, and is a compensatory mechanism for the accumulation of cytotoxic proteins and atrophy.
​
-
Greater dedifferentiation may include greater bilateral recruitment of neuronal resources as well as greater spread of local activation.
Methods - Data Collection
-
We conducted a cross-sectional study of 87 cognitively normal older adults
​
-
APOE E4 allele status was recorded for each participant.
​
-
Cerebral Aβ and glucose metabolism were measured with Pittsburgh Compound B (PiB) and fluorodeoxyglucose (FDG) tracers, respectively.
​
-
Participants were classified as PiB positive or PiB negative based on previous standard approaches [5].
​
-
Each participant’s cognitive performance was quantified using composite cognitive scores from neuropsychological assessments within the executive/attention, language, memory learning, memory retrieval, and visuospatial domains.
​​
-
During an fMRI scan, participants were shown face-name pairs and instructed to make a subjective judgement on whether each name “fit” the face shown above it
​
-
Face-name memory encoding task mimics everyday life and elicits reliable activation in our regions of interest [6].
Methods - Quantifying Dedifferentiation with Functional Asymmetry
-
Activation values were extracted from an fMRI scan during the face-name encoding task
​
-
Asymmetry index (AI) is measured as the laterality of the mean activation in each region of interest
​
-
A smaller absolute AI indicates greater dedifferentiation (i.e., greater bilateral activation).
Methods - Quantifying Dedifferentiation with Spread of Activation
-
To quantify the spread of neural activity, the peak activation voxels were first located at the four ROIs for all participants.
​
-
From the peak of the activation we approximated a gaussian curve emulating the drop of the activity as the region increases.
​
-
Greater FWHM indicates greater spread or dedifferentiation.
​​
Results - Associations between AD Factors and Activation Asymmetry
​
-
Multivariate regression in R looking for associations between activation asymmetry and all AD factors, controlling for age, education, race, and sex
​
-
Hippocampus AI was significantly explained with AD related factors (F(7,53) = 1.95, R2 = 0.205, p = 0.08)
​
-
PiB positive individuals showed more asymmetric activation (ß = 0.33, p = 0.024).
​​
Results - Associations between AD Factors and Activation Spread
​
-
Multivariate regression in R looking for associations between activation spread and all AD factors, controlling for age, education, race, and sex
​
-
Left hippocampus FWHM was significantly explained with AD related factors (F(7,53) = 2.07, R2 = 0.215, p = 0.063)
​
-
PiB positive individuals showed lower spread of activation (ß = -4.18, p = 0.019)
​​
Discussion and Characterization of Results​
-
PiB positive individuals were associated with greater asymmetry and lower spread of activation in the hippocampus.
​
-
Past studies have found that face-name encoding tasks generally involve symmetric activation, thus greater pathology may be associated with a loss of symmetry in this task.
​
-
Amyloid burden may overwhelm dedifferentiation mechanisms and hinder neuronal activation in the left hippocampus.
​
-
Greater asymmetry in PiB positive participants may be a product of lower overall activation and reduced spread of activation in the left hippocampus.
​
-
We found no association with cognitive function, which was not unexpected since these compensatory mechanisms may help to maintain cognitive function.
​​
