We have developed a way to map brain-wide networks using focal pulsed infrared neural stimulation in ultrahigh-field magnetic resonance imaging (MRI). The patterns of connections revealed are similar to those of connections previously mapped with anatomical tract tracing methods. These include connections between cortex and subcortical locations and long-range cortico-cortical connections. Studies of local cortical connections reveal columnar-sized laminar activation, consistent with feed-forward and feedback projection signatures. This method is broadly applicable and can be applied to multiple areas of the brain in different species and across different MRI platforms. Systematic point-by-point application of this method may lead to fundamental advances in our understanding of brain connectomes.

No abstract

Generalized approximate message passing (GAMP) is a promising technique for unknown signal reconstruction of generalized linear models (GLM). However, it requires that the transformation matrix has independent and identically distributed (IID) entries. In this context, generalized vector AMP (GVAMP) is proposed for general unitarily-invariant transformation matrices but it has a high-complexity matrix inverse. To this end, we propose a universal generalized memory AMP (GMAMP) framework including the existing orthogonal AMP/VAMP, GVAMP, and memory AMP (MAMP) as special instances. Due to the characteristics that local processors are all memory, GMAMP requires stricter orthogonality to guarantee the asymptotic IID Gaussianity and state evolution. To satisfy such orthogonality, local orthogonal memory estimators are established. The GMAMP framework provides a principle toward building new advanced AMP-type algorithms. As an example, we construct a Bayes-optimal GMAMP (BO-GMAMP), which uses a low-complexity memory linear estimator to suppress the linear interference, and thus its complexity is comparable to GAMP. Furthermore, we prove that for unitarily-invariant transformation matrices, BO-GMAMP achieves the replica minimum (i.e., Bayes-optimal) MSE if it has a unique fixed point.This paper is divided into two parts that can be read independently: The first part (main text) presents the universal framework of GMAMP, discusses the construction of BO-GMAMP and its properties. The second part (supplementary information) provides the additional explanatory for our main text including extended technical description of results, full details of mathematical models as well as all the proofs. The source code of this work is publicly available at sites.google.com/site/leihomepage/research.

No abstract

No abstract