“…Moreover, considering that, on occasion, drug concentration in blood, plasma, or CSF does not represent the concentration into the brain tissues or ISF, NDIs in the BBB and BCSFB may not significantly change its plasma concentration, whereas NDIs can affect the drug concentration in the brain and its efficacy against brain disorders [ 4 , 112 ]. Therefore, for the exact determination and prediction of NDIs, other kinds of measuring methods like brain imaging by using positron emission tomography and magnetic resonance spectroscopy are considered as a more adequate and relevant tool instead of drug quantification through blood or CSF collection, in recent times [ 16 , 113 ]. Moreover, several recent studies reported that deep learning models, kinds of in silico method for prediction of NDIs, can exhibit improved accuracy and more efficient performance, insisting that these models may play a key role in future research, drug discovery, and development processes to estimate possible interactions between natural compounds and new drugs for brain disorders [ 114 , 115 , 116 ].…”