Neurophysiological signals, sampling and instrumentation
A neurodiagnostic recording is the output of a measurement chain, not a direct photograph of every neural event. The signal depends on its source, sensor or electrode, geometry, amplifier, acquisition settings and noise. EEG measures electrical potential differences, whereas MEG measures magnetic fields. Evoked responses are interpreted relative to a stimulus and the recording arrangement. These modalities can provide complementary information without being interchangeable.
A differential amplifier measures the difference between inputs and is designed to reject common signals. Its performance also depends on electrode contact, impedance balance, cable arrangement and the actual equipment. A high nominal common-mode rejection ratio does not guarantee a clean patient recording. Investigate a noisy channel through the appropriate technical procedure rather than assume every high-amplitude event is cerebral.
Sampling and filtering affect what can be recorded and interpreted. The familiar twice-the-highest-frequency rule belongs to a band-limited sampling model; practical systems also require an appropriate margin and antialias filtering. Insufficient sampling can misrepresent higher-frequency activity at a lower apparent frequency. Once acquisition clips a waveform or loses information, changing the display scale does not restore the original signal.
Filters act on frequency content, not clinical meaning. A high-pass filter may reduce drift while altering a slow component. A low-pass or notch filter can affect transient shape, phase or timing according to its design. Preserve original data and document settings. A cleaner-looking trace is not automatically a more accurate representation of clinically relevant activity.
Averaging can improve a time-locked response relative to suitable random noise. The square-root relationship assumes conditions such as independent noise and a reasonably stable signal. Time-locked interference can persist, while latency jitter can blur the response. Reproducibility across independent runs is important, but even reproducible artifact can resemble a physiological component. Review morphology, distribution, timing and technical context together. This chapter supports conceptual acquisition reasoning; actual settings, electrical safety and device operation require the applicable validated procedure and trained staff.