Abstract Human functional magnetic resonance imaging (fMRI) typically employs the blood-oxygen-level-dependent (BOLD) contrast mechanism. In non-human primates (NHP), contrast enhancement is possible using monocrystalline iron-oxide nanoparticles (MION) contrast agent, which has a more temporally extended response function. However, using BOLD fMRI in NHP is desirable for interspecies comparison, and the BOLD signal's faster response function promises to be beneficial for rapid event-related (rER) designs. Here, we used rER BOLD fMRI in macaque monkeys while viewing real-world images, and found visual responses and category selectivity consistent with previous studies. However, activity estimates were very noisy, suggesting that the lower contrast-to-noise ratio of BOLD, suboptimal behavioural performance, and motion artefacts, in combination, render rER BOLD fMRI challenging in NHP. Previous studies have shown that rER fMRI is possible in macaques with MION, despite MION's prolonged response function. To understand this, we conducted simulations of the BOLD and MION response during rER, and found that no matter how fast the design, the greater amplitude of the MION response outweighs the contrast loss caused by greater temporal smoothing. We conclude that although any two of the three elements (rER, BOLD, NHP) have been shown to work well, the combination of all three is particularly challenging.
Introduction
In primate neuroimaging, researchers often face a methodological trade-off. Human fMRI studies rely almost exclusively on BOLD (Blood-Oxygen-Level-Dependent) contrast, which is non-invasive and offers a relatively fast hemodynamic response. Conversely, non-human primate (NHP) studies frequently utilize MION (monocrystalline iron-oxide nanoparticles), a contrast agent that provides a significantly higher signal-to-noise ratio (SNR) but possesses a much slower response function that lingers after stimulus cessation.
There is a strong desire to adopt rapid event-related (rER) designs in macaques to enable "condition-rich" experiments, such as those used in Representational Similarity Analysis (RSA). Theoretically, BOLD's faster temporal profile should make it superior for these high-speed designs. This study investigates whether the combination of rER designs and BOLD fMRI is viable in the macaque model.
Methodology
The study involved three rhesus macaques (M1, M2, M3) performing two types of fMRI tasks:
- Block-Design: Stimuli from specific categories (e.g., faces, places) were presented in sustained blocks to maximize signal detection.
- Rapid Event-Related (rER): 48 unique real-world images were presented for 0.5 seconds each, with a brief 2.5-second inter-stimulus interval.
Data were analyzed using univariate methods to localize category-selective regions and RSA to assess the discriminability of individual image patterns. Additionally, computational simulations compared the Impulse Response Functions (IRFs) of BOLD and MION to determine which contrast mechanism retains more information as experimental designs increase in speed.
Results
Success in Block Designs and Visual Mapping
The BOLD contrast proved effective during block-design experiments. The researchers successfully mapped broad visual responses across the cortex for all three subjects.
Furthermore, the block design allowed for the reliable localization of "face patches" within the temporal lobe, identifying regions with significantly higher activity for faces compared to places or objects. This confirmed that the BOLD signal in macaques is capable of identifying established functional landmarks when the experimental design provides sufficient power.
The Failure of rER BOLD
In contrast to the block design, the rER BOLD data were extremely noisy. While some basic category preferences were detectable at the group level, the signal was too weak to support fine-grained RSA. The researchers could not reliably distinguish the brain's response to one specific image versus another, a critical requirement for modern multivariate pattern analysis.
Simulation: Amplitude vs. Speed
The most critical finding came from computational simulations. While BOLD is "faster," MION's signal amplitude is so much larger that it provides a better SNR even in very rapid designs. In engineering terms, MION exhibits "full-spectrum dominance"; even after the high-frequency components of a rapid design are filtered through MION’s slow temporal response, the resulting signal remains stronger than the BOLD signal at those same frequencies.
Conclusion: The "Choose Two" Rule
The study concludes that researchers can successfully combine any two of the following three elements: rER designs, BOLD contrast, and NHP subjects. However, combining all three is exceptionally challenging. The inherent noise in NHP imaging—stemming from smaller brain volumes, motion artifacts, and behavioral variability—overwhelms the relatively weak BOLD signal in fast-paced designs.
For future NHP research, the authors recommend that if BOLD contrast is required for interspecies comparison, researchers should utilize block designs. If the research goals necessitate rapid, condition-rich designs, MION remains the superior and more reliable contrast mechanism due to its overwhelming power advantage.
Summary updated