Raman Spectroscopy (RS) and machine learning are explored for the rapid, real-time, sensitive application of neurological diagnostics through the human eye. Biochemical information is obtained of fatty porcine tissue and flat-mounted porcine retinal samples using an in-house built, portable RS system. RS and FUNDUS imaging have been combined with a phantom eye model to obtain spectra under eye-safe parameters in an in-vivo environment, identifying high wavenumber bands. This system has the potential to detect acute, biochemical changes indicative of neurodegenerative disorders such as Traumatic Brain Injury for early and accurate diagnoses, crucial for neurological recovery.
Near infrared spectroscopy (NIRS) has potential to offer a fast and non-invasive method of assessing cerebral
saturation in a clinical setting, however, there are concerns that NIRS brain measures suffer contamination
from superficial tissues. This study used the Valsalva manoeuver (VM) to determine whether NIRS could
differentiate between superficial (from somatic tissue) and neurological changes in the context of traumatic
brain injury. A potent vasopressor was used to assess the effect of reducing total haemoglobin concentration
in the superficial regions of the forehead. Frequency domain NIRS measurements during the VM pre and post
vasoconstrictor injection, combined with simulation data, conclusively show that NIRS can detect
neurological changes, in both haemoglobin content and saturation, when positioned on the forehead. The
effect of superficial contamination in this instance appeared to be insignificant, with no statistically significant
change in saturation over 8 patients, even with a drop in superficial haemoglobin concentration due to the
vasoconstrictor, confirmed by laser Doppler. Nevertheless, simulations indicated that the absolute values of
the recovered NIRS parameters are not quantitatively accurate; however a direct comparison with invasive
measures is needed to confirm this.
The subject of superficial contamination and signal origins remains a widely debated topic in the field of
Near Infrared Spectroscopy (NIRS), yet the concept of using the technology to monitor an injured brain,
in a clinical setting, poses additional challenges concerning the quantitative accuracy of recovered
parameters.
Using high density diffuse optical tomography probes, quantitatively accurate parameters from
different layers (skin, bone and brain) can be recovered from subject specific reconstruction models. This
study assesses the use of registered atlas models for situations where subject specific models are not
available. Data simulated from subject specific models were reconstructed using the 8 registered atlas
models implementing a regional (layered) parameter recovery in NIRFAST. A 3-region recovery based
on the atlas model yielded recovered brain saturation values which were accurate to within 4.6%
(percentage error) of the simulated values, validating the technique. The recovered saturations in the
superficial regions were not quantitatively accurate. These findings highlight differences in superficial
(skin and bone) layer thickness between the subject and atlas models. This layer thickness mismatch was
propagated through the reconstruction process decreasing the parameter accuracy.
Access to the requested content is limited to institutions that have purchased or subscribe to SPIE eBooks.
You are receiving this notice because your organization may not have SPIE eBooks access.*
*Shibboleth/Open Athens users─please
sign in
to access your institution's subscriptions.
To obtain this item, you may purchase the complete book in print or electronic format on
SPIE.org.
INSTITUTIONAL Select your institution to access the SPIE Digital Library.
PERSONAL Sign in with your SPIE account to access your personal subscriptions or to use specific features such as save to my library, sign up for alerts, save searches, etc.