Abstracts

Rationale: Focal brain cooling, which aims at decreasing the temperature of epileptogenic cortex, has attracted attention as a potential treatment for epilepsy. Animal studies have shown that it can suppress the focal seizures (Epilepsia 2012; 53(3); 485-493). Hence, we are developing a focal brain cooling system for clinical application. The system delivers cooled physiological saline to a cooling device which is made of titanium and implanted in the skull. The outflow is cooled by a thermoelectric device and pumped for circulation. The thermoelectric device and the pump are driven by a mobile battery, which is activated only when the seizures are predicted. To realize this system, in this research, numerical simulations on the cooling performance were performed to calculate the time required for cooling the epileptogenic cortex and to investigate the influence of the channel structure in the cooling device. Methods: Numerical simulations of mass and heat transfer were conducted with a human head model. The geometrical representation of the model is depicted in Fig. 1 (Eur J Appl Physiol 1998; 78(4); 353-359). The model of soft tissue, skull, cerebrospinal fluid (CSF), and brain consisted of the conventional bioheat equation proposed by Pennes (J Appl Physiol 1948; 1(2); 93-122), and the mass and heat transfer in the cooling device and the saline, flowing inside the tube in the soft tissue, was calculated by computational fluid dynamics (CFD) simulation. The flow rate and inflow temperature of saline were 400 mL/min and 10 C, respectively. Under these conditions, unsteady and steady simulations were conducted. Through the unsteady simulations, the length of time was calculated for the average and maximum temperatures on target region of brain surface decreasing from 38 C to 15 C by the cooling device of type D in Table 1. Through the steady simulations, four types of channel structure in the cooling device were designed and temperature distribution on the target region and pressure drop of the saline were calculated. Results: The results of the unsteady simulations for type D revealed that the average and maximum temperatures on the target region reached 15 C in less than 5 and 34 seconds, respectively. These are much shorter than the average length of time from prediction to occurrence of the epileptic seizures when an existing algorithm for seizure prediction was used (Neurophysiol Clin 2005; 116; 532-544). The results of the steady simulations, shown in Table 1, revealed a trade-off between cooling performance and pressure drop, which affects the electrical consumption of battery. Conclusions: The results of the unsteady simulations confirmed that the time required for cooling the epileptogenic cortex is shorter than the time from prediction to occurrence of the epileptic seizures. The steady simulations enable us to investigate the trade-off between the cooling performance and the electrical consumption for designing the whole cooling system. Funding: This study was supported by JSPS KAKENHI Grant Number 15H05719.