Existing brain imaging techniques, such as fMRI and PET, severely restrict movement, hindering research on populations with movement disorders (e.g., Parkinson's Disease, dementia) and ecologically valid studies of natural behaviors. While techniques like EEG and NIRS allow for upright movement, their resolution and brain penetration are limited. PET offers quantitative measurement of metabolite uptake, making it ideal for studying pharmaceutical effects, but its motion sensitivity poses a significant challenge. Previous attempts to address motion tolerance in PET include repeated injections of O15-H2O, which exposes participants to high radiation, and delayed-PET studies, which require a separate baseline scan and limit the number of tasks performed in a single session. This study aims to overcome these limitations by using a novel wearable motion-enabled PET imager, the AMPET helmet system, to image brain activity during a walking-in-place task. The AMPET system, weighing less than 3 kg, is suspended from above, allowing comfortable head movement. This study validates the AMPET prototype by testing for motion artifacts and assessing task-related activity in motor-related brain regions.

The introduction extensively reviews existing neuroimaging techniques and their limitations regarding motion tolerance and brain coverage. It highlights the need for a system that allows for both deep brain imaging and natural upright movement. The limitations of fMRI, PET, SPECT, EEG, NIRS, and HD-DOT are discussed in detail. The review also covers previous attempts to adapt PET for motion tolerance, such as using O15-H2O and delayed-PET methods. The advantages and disadvantages of each approach are analyzed, setting the stage for the introduction of the AMPET system as a potential solution.

This study used a convenience sample of 11 cancer patients (mean age 53, range 32–76) already scheduled for a clinical PET scan. Participants performed a walking-in-place task while wearing the AMPET helmet after receiving a bolus injection of 18F-FDG. The AMPET prototype consists of 12 detector modules with a 21 cm FOV and 2 mm spatial resolution. The imager was suspended from above to minimize the weight felt by the participant. The study used three validation measures: (1) Assessing motion artifacts during walking vs. standing still; (2) Validating activation in a priori defined cortical ROIs related to movement; and (3) Validating differential activation in deep brain structures like the thalamus and basal nuclei. Image analysis involved co-registration of AMPET scans with clinical PET/CT scans using MIM software. ROIs were defined based on a previous PET study of walking. Image intensity was normalized using a reference ROI in the frontal cortex. Statistical analyses included mixed-effects models and t-tests. One participant's data was omitted due to insufficient imaging time, and two participants' data were excluded due to incorrect helmet placement. A subset of participants (n=4) was reimaged with a lower helmet placement to assess deep brain structures.

Motion artifact analysis showed no significant differences in whole-brain activity to FOV ratio between walking and standing still, indicating that the AMPET system tolerated motion with minimal artifact (<1.3 mm misalignment). Functional analysis revealed significantly greater activity in the bilateral leg-M1 ROI during walking compared to other ROIs (p<0.001), consistent with previous studies. Activity in the SMA was significantly greater than other non-motor cortical regions but less than the leg-M1 ROI. In a single amputee participant, greater activity was observed in the right hemisphere leg-M1 ROI (representing the intact leg), which is congruent with similar findings. Analysis of deep brain structures in a subset of participants revealed a trend towards greater activation in the caudate nucleus compared to other regions during walking, supporting the feasibility of deep brain imaging with AMPET. Re-imaging of the leg-M1 ROI in a subset of participants at a later time point showed consistent activation, validating the stability of the method.

This study demonstrates the feasibility of performing ecologically valid PET neuroimaging of the brain during upright ambulatory motor tasks using the AMPET system. The minimal motion artifacts observed indicate the system's robustness, and the functional results support the system's ability to detect task-related activity in both cortical and deep brain structures. The findings have implications for research on populations with movement disorders and for studies of complex real-world behaviors. The inclusion of an amputee participant yielded insights into motor cortex changes in individuals with limb loss. The study highlights the need for future developments in AMPET, such as incorporating an optical positioning system to improve image quality and enhance deep brain coverage.

The AMPET prototype successfully demonstrated the feasibility of real-time, motion-tolerant PET neuroimaging during upright human walking. The study validated its performance by confirming the presence of motion-related activity in expected cortical and subcortical regions. Future improvements, including an optical positioning system and expanded FOV, will further enhance the utility of this technology for a wide range of ecologically valid neuroscience studies.

The study used a small convenience sample of cancer patients, which may limit the generalizability of the findings. The use of a single bolus injection without arterial blood sampling might have affected the accuracy of quantitative measurements. The simplified suspension system and the limited FOV of the prototype AMPET imager also pose limitations, as does the small number of participants in the deep brain imaging analysis.