Precise acoustic field establishment by holography-modulated acoustic intensity

Acoustic holography, which reconstructs desired target acoustic fields by precisely controlling the phase distribution of acoustic wavefronts, holds significant promise for applications such as acoustic manipulation. However, the precise modulation of acoustic field distributions via acoustic holography to construct multifocal fields with controllable acoustic intensity ratios remains insufficiently explored. To address this limitation, this study proposes a Physics-Informed Artificial Intelligence-based Angular Spectrum method (AIAS), which deeply integrates the physical model of angular spectrum propagation into the neural network training process. Combined with a specifically designed Target-Area-Weighted Mean Squared Error loss function, AIAS establishes an explicit optimization link between the phase distribution and the amplitude error in the target region during the inverse design process. Results demonstrate that acoustic fields reconstructed by AIAS exhibit more concentrated and uniform pressure distributions (average pressure improved from 262 ± 15 kPa to 276.4 ± 5.6 kPa), providing stable acoustic fields for particle assembly. Importantly, by controlling the phase gradient distribution, AIAS successfully constructs asymmetric acoustic fields with a 2:1 intensity ratio between two focal points. The exceptional amplitude modulation capabilities of AIAS represent a key technological breakthrough for achieving more precise and personalized transcranial focused ultrasound therapy.