Skull Impact on Photoacoustic Imaging of Multi-Layered Brain Tissues with Embedded Blood Vessel Under Different Optical Source Types: Modeling and Simulation

Xi Yang, Chengpeng Chai, Yun-Hsuan Chen, Mohamad SawanBioengineering 2025 · DOI

TL;DR

This study investigates how the human skull's optical and acoustic properties affect photoacoustic imaging (PAI) of brain structures, specifically blood vessels. By simulating PAI with various light sources and accounting for the skull's presence, the researchers found that while initial signal strength is consistent across some sources, the skull significantly alters reconstructed images, impacting vessel detection and morphology. This work provides a crucial theoretical foundation for optimizing PAI for human brain applications.

What Problem Does This Paper Address?

Photoacoustic imaging (PAI) holds immense promise for non-invasively visualizing biological tissues due to its ability to combine optical specificity with acoustic penetration depth. However, applying PAI to the human brain is severely hampered by the skull. The skull acts as a formidable barrier, exhibiting high optical scattering and significant acoustic attenuation and distortion. This makes it incredibly difficult to generate and detect clear photoacoustic signals from deep within the brain. While previous research has explored the optical challenges, understanding the combined impact of optical and acoustic phenomena, especially in conjunction with different illumination strategies, remains a critical hurdle for translating PAI to human brain imaging. This paper aims to bridge this gap by simulating PAI of a multi-layered brain model with an embedded blood vessel, systematically evaluating the influence of the skull and various optical source types on image reconstruction.

Methods

The researchers employed a multi-stage simulation approach:

  1. Brain Model: A simplified four-layer model representing the scalp, skull, cerebrospinal fluid (CSF), and gray matter was constructed. A 2mm diameter blood vessel was embedded within the gray matter. The model dimensions were 14x14x14 mm³, with a grid resolution of 0.1 mm.
  2. Optical Simulation (Monte Carlo): The mcxlab toolbox was used to simulate light propagation through the brain model. Sixteen different optical source types (e.g., pencil, isotropic, Gaussian, planar, disk) were tested. The simulation calculated the optical fluence, from which optical absorption was derived using the Beer-Lambert law.
  3. Acoustic Simulation (k-wave): The calculated optical absorption served as the initial pressure source for acoustic simulations using the k-wave toolbox. Two acoustic models were used:
  4. Acoustic-Homogeneous: Assumed uniform acoustic properties (speed of sound and density) similar to soft tissue/water.
  5. Acoustic-Heterogeneous: Incorporated the specific acoustic properties of the skull (higher speed of sound and density) while using water-like properties for other layers.
  6. Image Reconstruction (Delay and Sum): The simulated acoustic sensor data was processed using the Delay and Sum (DAS) beamforming algorithm to reconstruct photoacoustic images.
  7. Image Evaluation: The reconstructed images were analyzed for signal distribution in 2D, horizontal, and vertical directions. Comparisons were made between simulations with and without skull properties, and across the sixteen different optical source types. Maximum and minimum PA signal values and their corresponding source types were identified.

Datasets

Not discussed in the paper. The study relies on simulated data generated from the defined brain model and optical/acoustic simulation parameters.

Key Insights and Ideas

  • Skull's Dual Threat: The skull isn't just an optical hurdle; its acoustic properties significantly distort and attenuate the photoacoustic signals, making reconstruction challenging.
  • Source Matters, But How? While the initial photoacoustic signal strength from different optical sources shows some consistency in terms of which sources are strongest/weakest, the reconstructed images reveal a different story. The skull's influence causes the optimal optical source type for imaging to change.
  • Morphological Distortion: The presence of the skull can alter the apparent shape and characteristics of embedded structures like blood vessels, turning them into inverted half-circular shapes in some cases, which can complicate interpretation.
  • Signal Attenuation is Pervasive: Both optical scattering and acoustic attenuation lead to a general decrease in signal intensity, particularly in deeper regions, and can introduce artifacts.
  • Acoustic Simulation is Crucial: Simply considering optical propagation is insufficient. The acoustic simulation, especially with heterogeneous skull properties, is vital for understanding how signals are transmitted and how they impact the final reconstructed image.

Key Results

  • Initial PA Signals: For the initial PA signals, the strongest and weakest signals at different positions were consistently observed from specific optical source types (e.g., pencil array, collimated Gaussian, planar for maximum; pencil, cone, angular Gaussian for minimum).
  • Reconstructed Images (Without Skull): Reconstructed images without the skull showed more distinct layer characteristics and a clearer representation of the blood vessel.
  • Reconstructed Images (With Skull): The presence of the skull led to signal distortion, reduced intensity, and artifacts. The blood vessel appeared as inverted half-circular shapes with strong signals in the middle.
  • Optical Source Performance Varies: The ranking of optical source types for optimal imaging performance changed significantly between initial PA signals and reconstructed images, especially when the skull was included. For instance, while pencil array was strong initially, its performance in reconstructed images with the skull was less dominant.
  • Depth Dependence: PA signals generally decreased with depth due to attenuation. The skull's influence was most pronounced at the boundaries between tissue layers.

Limitations and Future Work

  • Simplified Model: The study used a simplified four-layer brain model. A more complex, anatomically accurate model could provide richer insights.
  • Grid Size: The 0.1 mm grid size, while suitable for optical simulations, might be a simplification for acoustic simulations where wavelength-dependent resolutions are critical.
  • Linear Array: The use of a linear ultrasound array can introduce artifacts. Future work could explore different array geometries (e.g., circular, spherical) to mitigate this.
  • Computational Cost: The simulations are computationally intensive. Developing more efficient algorithms or leveraging advanced computational platforms could enable larger-scale studies.
  • Experimental Validation: The study is purely simulation-based. Future work should focus on experimental validation to confirm these findings in real biological tissues.
  • Deep Learning: The authors mention deep learning methods for improving image quality in the introduction, suggesting this could be a future avenue for artifact reduction and signal enhancement in skull-affected PAI.
  • Optimized Source Design: The study provides a theoretical basis for selecting optical sources. Future research could focus on designing novel optical sources tailored to overcome skull-induced challenges.

Summary updated

Abstract

Skulls with high optical scattering and acoustic attenuation are a great challenge for photoacoustic imaging for human beings. To explore and improve photoacoustic generation and propagation, we conducted the photoacoustic simulation and image reconstruction of the multi-layer brain model with an embedded blood vessel under different optical source types. Based on the optical simulation results under different types of optical sources, we explored the characteristics of reconstructed images obtained from acoustic simulations with and without skull conditions. Specifically, we focused on the detection of blood vessels and evaluated the image reconstruction features, morphological characteristics, and intensity of variations in the target vessels using optical and acoustic simulations. The results showed that under the initial PA signals, the types of optical source types corresponding to the strongest and weakest photoacoustic signals at different positions within the target region were consistent, while the optical source types were different in the reconstructed images. This study revealed the characteristics of acoustic signal transmission with and without skull conditions and its impact on image reconstruction. It further provides a theoretical basis for the selection of optical sources.
BibTeX
@inproceedings{yang2025skull,
  author = {Xi Yang and Chengpeng Chai and Yun-Hsuan Chen and Mohamad Sawan},
  title = {{Skull Impact on Photoacoustic Imaging of Multi-Layered Brain Tissues with Embedded Blood Vessel Under Different Optical Source Types: Modeling and Simulation}},
  booktitle = {Bioengineering},
  year = {2025},
  doi = {10.3390/bioengineering12010040},
}
Figure