Synopsis

Keywords: Neuro: Brain function, Image acquisition: Artefacts

Functional MRI signals are far from "clean"! Along with actual random noise, the signals can be affected by artifacts and confounds that are “systematic” – in other words, with characteristic structure in space and/or in time. These can be linked to head motion, cardiac and respiratory activity, and imperfections in the imaging process, for example. Understanding these phenomena can be critical to avoid biased results and interpretations. In this educational talk, I will cover the diverse sources of systematic noise that we currently know about, including their origins and mechanisms, properties, and their potential impact on fMRI studies.

Syllabus

Functional MRI signals are far from "clean"! Along with actual random noise, the signals can be affected by artifacts and confounds that are “systematic” – in other words, with characteristic structure in space and/or in time. Their sources are diverse: some originate from the subject (e.g. head motion, cardiac and respiratory activity, cerebrospinal fluid fluctuations), while others originate from the imaging process (e.g. parallel imaging artifacts, imperfect sampling and hardware limitations). Due to their structured properties, if overlooked or underestimated, these interferences can sneak up on us and confound our analyses, posing as “authentic” brain activity – and significantly bias our interpretations of what the brain is actually doing. In this educational talk, I will cover the diverse sources of systematic noise that we currently know about, describing fundamental aspects such as:

• Their origins and mechanisms;
• Spatial, temporal, spectral properties;
• Dependence on field strength and acquisition strategies;
• Their potential impact on fMRI studies (resting-state, task-based).

This knowledge should prove valuable to help researchers understand systematic noise and its effects, and provide a useful basis (and motivation) to subsequently learn about existing strategies for their prevention and correction. Altogether, this knowledge should help us study the brain more effectively, with less biased, better-informed interpretations.

Acknowledgements

I gratefully acknowledge funding from the Swiss National Science Foundation through grant 185909, and the support of CSEM – Swiss Center for Electronics and Microtechnology.

References

No reference found.