This paper presents a time-domain based lossless data reduction technique called Log2 Sub-band encoding, which is designed for reducing the size of data recorded on a wireless electroencephalogram (EEG) recorder. A data reduction unit can help to save power from the wireless transceiver and from the storage medium since it allows lower data transmission and read/write rates, and then extends the life time of the battery on the device. Our compression ratio(CR) results show that Log2 Sub-band encoding is comparable and even superior to Huffman coding, a well known entropy encoding method, whilst requiring minimal hardware resource, and it can also be used to extract features from EEG to achieve seizure detection during the compression process. The power consumption when compressing the EEG data is presented to evaluate the system0s overall improvement on its power performance, and our results indicate that a noticeable power saving can be achieved with our technique. The possibility of applying this method to other biomedical signals will also be noted.