Introduction There are many cases of multilevel lumbar spinal stenosis and foraminal stenosis, but there is still no neuroelectrophysiological diagnostic method with a high spatial resolution for these conditions. We have been developing a magnetospinograph system which is a supine position-type biomagnetometer equipped with superconducting quantum interference device (SQUID) fluxmeter. We have reported that magnetospinography could noninvasively evaluate the spinal cord and spinal nerve activity by body surface recordings. Methods We used a newly developed 124-channel SQUID biomagnetospinometer system (Kanazawa Institute of Technology, Japan) with a wide sensor area and a suitable curve for lumber lordosis. Neuromagnetic fields of 32 healthy subjects (64 nerves), 22–58 years of age (mean age 41 years), were measured at the surface of the lower back after electric stimuli of the peroneal nerves at the knees. With the use of a unit gain normalized minimum-norm (UGMN) filter method, which is a spatial filtering technique, the direction and intensity of the current source on each reconstruction point was reconstructed, and the conductive direction, distribution, and temporal change of the current source were visualized. Then, the propagation pathway of the estimated electrical currents was superimposed on lumbar spine X-rays. Results We used a newly developed 124-channel SQUID biomagnetospinometer system (Kanazawa Institute of Technology, Japan) with a wide sensor area and a suitable curve for lumber lordosis. Neuromagnetic fields of 32 healthy subjects (64 nerves), 22–58 years of age (mean age 41 years), were measured at the surface of the lower back after electric stimuli of the peroneal nerves at the knees. With the use of a unit gain normalized minimum-norm (UGMN) filter method, which is a spatial filtering technique, the direction and intensity of the current source on each reconstruction point was reconstructed, and the conductive direction, distribution, and temporal change of the current source were visualized. Then, the propagation pathway of the estimated electrical currents was superimposed on lumbar spine X-rays. For all subjects, we could record the nerve action magnetic fields in the lumbar spine after peroneal nerve stimuli. In 44 nerves (68.8%), we could visualize the estimated electrical currents propagating through the surrounding fifth intervertebral foramina and ascending to the cranial direction in the lumbar canal. The conduction velocity of the current source was 40.3–78.8 m/s (mean 56.5 m/s), as calculated from the peak latency. Conclusion Using the magnetospinograph system, we successfully visualized the propagation of estimated neural electric activities in the lumbar spine after peroneal nerve stimuli. This system holds enormous potential for establishing a non-invasive neuroelectrophysiological diagnostic method for the identification of the disordered site in lumbar spinal nerves.