Computer-guided optical clearing for transcranial laser speckle imaging of cortical blood flow through synergistic tartrazine-induced cranial bone transparency

The clinical deployment of transcranial laser speckle-contrast imaging is limited by pronounced multiple scattering in cranial bone, which substantially reduces probing depth and distorts estimates of cerebral blood flow parameters. In the present study, a hybrid strategy combining computational optical clearing with tartrazine induced optical clearing of cranial bone is proposed and experimentally validated. Topical application of a 30% (w/v) tartrazine solution reduces the scattering coefficient of cranial bone from ∼1.5 to ∼0.6 mm-1 by refractive index matching and by decreasing the intratissue heterogeneity of their spatial distribution. Residual quasi static scattering components persisting after tartrazine induced optical clearing are selectively removed by computational optical clearing based on principal component analysis filtering, yielding an ≂ two-fold increase in vessel contrast-to-noise ratio (from 1.5 to 3.1) and exposing vascular networks invisible with conventional transcranial LSCI. This algorithm automatically ranks components in descending order of their contribution to the total variance: components with large eigenvalues correspond to quasi static structures (cranial bone and stationary brain tissue), whereas those with small eigenvalues contain the dynamic signal arising from blood flow. Applying the Guttman-Kaiser criterion enables selective reconstruction of the dynamic component while suppressing background scattering. This synergistic approach opens new possibilities for noninvasive, high-contrast visualization of cerebral blood flow and represents a promising, qualitatively new direction in the development of physico-digital optical clearing technologies.