The Best SD Distribution Of Noninvasive Detection Of Cranial Dural Hematoma In Near-infrared Spectroscopy

Due to the advantages of rapid and non-invasive detection of traumatic dural hematoma using near-infrared spectroscopy method,this technology has become a hot research topic in tissue optics in recent years and has important applications in clinical emergency treatment.Rapid diagnosis of traumatic brain injury in complex environment such as natural disaster is the key to reduce the mortality of dural hematoma.Near infrared spectroscopy is one of the methods for rapid diagnosis of cerebral hematoma,at present,there are some related research results,but the study on the optimal SD distribution of the brain dural hematoma detected by near-infrared spectroscopy has not been published.The location and number of detectors around the light source are critical for reducing the pathological characteristics and complexity of the prediction model on dural hematoma degree,besides,It can provide more abundant reference information for the basic research of brain function.To further improve the detection accuracy of dural hematoma detection,in this paper,a multi-channel differential optical density method was used to obtain the bilaterally-symmetric optical density data of brain.Therefore,rational selection of detector numbers and their distances from the light source is very important.The main contents include:Recording optical density of 2.5cm,3.0cm,3.5cm,4.0cm,4.5cm,in the case of different hematoma with Monte Carlo(MC).The equivalent signal to noise ratio concept is proposed for the measurement of the optimal S-D distribution,and the optimal S-D distribution is simulated by using 808nm reference light,and a detector position screening method based on Variable Importance in the Projection(VIP)analysis is proposed,in the end,validation the optimum S-D using the resin.In this paper,the model correlation of the dural hematoma degree is 88.94%,average error is 0.0838cm-1,the prediction model correlation of sample number 16 is 86.65%,average error is 0.1047cm-1.The optimal S-D range is obtained according to the ESNR.To establish optimal S-D prediction model for different populations 808nm,near infrared light is used as reference light.The model correlation of reference light training set is 99.1%,average error is 0.0291cm,the prediction model correlation of sample number 16 is 98.8%,average error is 0.0323cm,After VIP analysis,the number of detectors was reduced from 30 to 4(2.1cm,2.4cm,3.4cm,4.2cm)and the MRE of the dural position ?a prediction was reduced from 4.08%to 2.06%after the reduction in detector numbers.The experiments are carried out with 4 kinds of tissue imitation.The results verify the research about the optimal S-D distribution,also showed that the dural hematoma patients cannot be accurately diagnosed with single detector.