| Summary: | Diffuse optical tomography (DOT) was investigated as a means of non-invasively monitoring neonatal brain function at the cot-side. Two DOT devices designed and built at UCL were used in this work: a continuous-wave (CW-DOT) system and a time-resolved (TR-DOT) system known as MONSTIR II. Both systems are sensitive to changes in absorption within the interrogated tissue and can be used to create images of relative changes in oxy- and deoxyhaemoglobin concentration. MONSTIR II also has the capacity to isolate changes in absorption from those in scattering by obtaining photon flight-time information. These DOT devices were used to explore whether infant neuropathologies, such as perinatal arterial ischaemic stroke (PAIS), can be detected earlier than is currently achieved clinically. To test the efficacy of DOT for early PAIS diagnosis, stroke-affected infants were initially studied with a CW-DOT system. Significant differences in interhemispheric correlation were found between the control group and the PAIS group, with the latter having impaired connectivity in the frequency range of 0.0055-0.080 Hz. The efficacy of MONSTIR II for imaging similar neuropathological infants was evaluated using a complex anatomically-accurate neonatal head phantom created using a novel 3D printing technique. To experimentally determine the optimal combination of near-infrared wavelengths for studying brain function with MONSTIR II, a widely applicable systems-testing protocol based on functional activation was developed. To mitigate the inherently slow sampling of TR-DOT systems, a variational form Kalman filter (VKF) was implemented to improve the spatio-temporal resolution of MONSTIR II for imaging transient neuropathological events. The VKF was tested on two dynamic phantoms designed to mimic neonatal seizures. Finally, MONSTIR II was used to acquire data from healthy and brain injured infants. Preliminary results from a neonate with PAIS suggest a significant asymmetry between hemispheres for wavelength intensity ratio data.
|
|---|
