Magnetoencephalography with a Cs-based high-sensitivity compact atomic magnetometer.

Abstract

In recent years, substantial progress has been made in developing a new generation of magnetoencephalography (MEG) with a spin-exchange relaxation free (SERF)-based atomic magnetometer (AM). An AM employs alkali atoms to detect weak magnetic fields. A compact AM array with high sensitivity is crucial to the design; however, most proposed compact AMs are potassium (K)- or rubidium (Rb)-based with single beam configurations. In the present study, a pump-probe two beam configuration with a Cesium (Cs)-based AM (Cs-AM) is introduced to detect human neuronal magnetic fields. The length of the vapor cell is 4 mm, which can fully satisfy the need of designing a compact sensor array. Compared with state-of-the-art compact AMs, our new Cs-AM has two advantages. First, it can be operated in a SERF regime, requiring much lower heating temperature, which benefits the sensor with a closer distance to scalp due to ease of thermal insulation and less electric heating noise interference. Second, the two-beam configuration in the design can achieve higher sensitivity. It is free of magnetic modulation, which is necessary in one-beam AMs; however, such modulation may cause other interference in multi-channel circumstances. In the frequency band between 10 Hz and 30 Hz, the noise level of the proposed Cs-AM is approximately 10 f T/Hz1/2, which is comparable with state-of-the-art K- or Rb-based compact AMs. The performance of the Cs-AM was verified by measuring human auditory evoked fields (AEFs) in reference to commercial superconducting quantum interference device (SQUID) channels. By using a Cs-AM, we observed a clear peak in AEFs around 100 ms (M100) with a much larger amplitude compared with that of a SQUID, and the temporal profiles of the two devices were in good agreement. The results indicate the possibility of using the compact Cs-AM for MEG recordings, and the current Cs-AM has the potential to be designed for multi-sensor arrays and gradiometers for future neuroscience studies.