Iron is an essential nutrient in the human body, including the brain-- especially during neurodevelopment. Both iron deficiency and iron overload present significant risks to the development and function of the young brain. Iron in the brain can be detected using magnetic resonance (MR) imaging and is visible on T2*-weighted gradient-echo (GRE) images as hypointensity caused by field heterogeneity and magnetic susceptibility effects [1]. R2* is the transverse relaxation rate measured directly from multiecho GRE MR images acquired at different echo times and is influenced by iron content. Increase in brain iron is part of normal brain development, and has been observed to have maximal concentration in the globus pallidus (GP) [2]. In this study we investigate iron deposition in the GP of healthy youth between the ages of 8 and 22.

Imaging protocol: Data was acquired in a subsample of 815 typically developing subjects from the Philadelphia Neurodevelopmental Cohort (PNC) on a 3T MR scanner (Magnetom Trio, Siemens, Erlangen, Germany) using a 32-channel receive-only head coil. The magnitude data from a multiecho GRE sequence with the following parameters was used to generate T2* images based on equation (1): resolution of 3.8 x 3.8 mm2, 4mm slice thickness, no slice gap, TE1/TE2/TR=2.69ms/5.27ms/1000ms, 44 axial slices.

T2* = -△TE / ln (I_TE2/I_TE1)...(1)

Data Analysis: Each subject’s T2* image was registered to their respective high resolution T1, MPRAGE images and transformed to MNI space. Mean T2* was measured from the right and left Pallidum using the respective ROIs from the Harvard-Oxford subcortical structural atlas. R2* was calculated as the inverse of T2*, R2* = 1/T2* (1/sec). Right and left R2* values were combined as no significant hemispheric differences in R2* were observed. R2* values greater than and less than two standard deviations from the mean were excluded (n=51). The mean R2* values were plotted by age and fitted to the monoexponential model described in [2], Figure 2. Mean R2* values were analyzed using the Generalized Linear Model (GLM) with age and sex as predictors. Follow-up GLM analysis was performed by distinct age bins (prepubertal-under age 12), pubertal transition (age 12-16), and adult (postpubertal, age 17-22) into three age groups. Iron deposition (R2*) (Mean ± SE) was estimated for each gender by group using 'lsmeans' (least squares means) package in statistical software package R with the p values adjusted using the tukey method for 2 means.

The mean R2* in the GP was measured as 21.87 sec-1 with a standard deviation of 1.88 sec-1 and is comparable to previously reported values [3]. Overall, iron deposition was significantly associated with age (p < 0.001) and sex (p < 0.05). Younger individuals had less iron deposition as compared to older individuals and males had higher iron deposition as compared to females. Follow-up analysis of the R2* values with specific age ranges resulted in a significant relationship between iron deposition and age group (p < 0.001) and a significant age by sex interaction (p < 0.01). Males in the pubertal transition group had higher iron deposition (p < 0.001) within the GP as compared to females (Figure 3). No significant gender differences were present in the prepubertal or postpubertal group.Iron levels in the GP estimated using R2* are significantly correlated with age in children and young adults. The temporal sequence of iron deposition is in line with previously reported work. Iron levels in the brain of pubertal girls are significantly lower than iron levels in pubertal boys. The gender difference in iron levels we observe in the pubertal transition group may be related to the high prevalence of iron deficiency in pubertal girls due to menstrual blood loss, lifestyle, and dietary habits [4].