1: \begin{abstract}

2: 	We consider the problem of precision matrix estimation where, 

3: 	due to extraneous confounding of the underlying precision matrix, 

4: 	the data are independent but not identically distributed. While such confounding occurs in many scientific problems, 

5: 	our approach is inspired by recent neuroscientific research suggesting that brain function, as measured using functional magnetic resonance imaging (fMRI), 

6: 	is susceptible to confounding by physiological noise, such as breathing and subject motion. 

7: 	Following the scientific motivation, we propose a graphical model, which in turn motivates 

8: 	a joint nonparametric estimator.  We provide theoretical guarantees for the consistency 

9: 	and the convergence rate of the proposed estimator. In addition, we demonstrate that the

10: 	optimization of the proposed estimator can be transformed into a series of 

11: 	linear programming problems and, thus, can be efficiently solved in parallel.

12: 	Empirical results are presented using simulated and real brain imaging data and

13: 	suggest that our approach improves precision matrix estimation as compared to baselines 

14: 	when confounding is present.

15: \end{abstract}

16: