Despite application of brain-computer interfaces (BCI) in many fields, there is little research on the use of BCI to identify sensory signals in the human brain. This SW Project aimed to identify different brain wave patterns under two basic needs: hunger and thirst; and two physiological states: cold and heat. The goal was to assist individuals who are brain active but unable to communicate their needs due to neural injuries or disorders, such as someone with lock-in syndrome or a coma. Using a non-invasive OpenBCI soft electrode cap, the project collected brain wave signals from participants experiencing 2 basic needs: hunger, and thirst; and 2 physiological states: cold and heat. Collaboratively designed, the experiment had three phases: baseline brain wave measurements of people in normal conditions, measurements of brain wave patterns during hunger or thirst, and those during feeling hot or cold. The brain wave patterns in the five states showed differences after visualization. We used a combined machine learning approach on this data to teach an EEGConformer model to identify participants’ states. After around 200 epoch, the final model found significant differences between treatments and achieved an accuracy rate of test datasets at about 83%.