Model-based navigation of transcranial focused ultrasound neuromodulation in humans: Application to targeting the amygdala and thalamus

Abstract

Background: Transcranial focused ultrasound (tFUS) neuromodulation has shown promise in animals but is challenging to translate to humans because of the thicker skull that heavily scatters ultrasound waves.

Objective: We develop and disseminate a model-based navigation (MBN) tool for acoustic dose delivery in the presence of skull aberrations that is easy to use by non-specialists.

Methods: We pre-compute acoustic beams for thousands of virtual transducer locations on the scalp of the subject under study. We use the hybrid angular spectrum solver mSOUND, which runs in ∼4 s per solve per CPU yielding pre-computation times under 1 h for scalp meshes with up to 4000 faces and a parallelization factor of 5. We combine this pre-computed set of beam solutions with optical tracking, thus allowing real-time display of the tFUS beam as the operator freely navigates the transducer around the subject' scalp. We assess the impact of MBN versus line-of-sight targeting (LOST) positioning in simulations of 13 subjects.

Results: Our navigation tool has a display refresh rate of ∼10 Hz. In our simulations, MBN increased the acoustic dose in the thalamus and amygdala by 8-67 % compared to LOST and avoided complete target misses that affected 10-20 % of LOST cases. MBN also yielded a lower variability of the deposited dose across subjects than LOST.

Conclusions: MBN may yield greater and more consistent (less variable) ultrasound dose deposition than transducer placement with line-of-sight targeting, and thus could become a helpful tool to improve the efficacy of tFUS neuromodulation.

Keywords: Acoustic modeling; Hybrid angular spectrum; Low intensity focused ultrasound pulsation (LIFUP); Neuromodulation; Neuronavigation; Transcranial focused ultrasound (tFUS).

Fig. 1.

Pre-calculation GUI. Panel 1: The operator loads and visualizes the head mask, bone porosity and segmentation index map of deep nuclei structures (‘aseg’ map from FreeSurfer/SAMSEG). Toggling between those maps is useful to check for possible alignment errors. Panel 2: A mesh of the scalp is generated and visualized. The operator can adapt the number of faces in the mesh by changing the average triangle edge length and removing faces that do not correspond to valid transducer positions (e.g. below the ears). Faces’ normal are displayed to verify their proper orientation. Panel 3: Details of the transducer modeled. Panel 4: As the computation proceeds, mSOUND computes beam solutions corresponding to the transducer positions/faces of the scalp mesh in parallel. A typical pre-calculation time is 30 min for 4000 mesh faces with a parallel factor 10.

Fig. 2.

Navigation GUI. The operator loads the results of the pre-calculation, selects the target brain nucleus and can save the various display panels as Matlab figures. The current implementation of the tool uses the so-called ‘aseg’ map generated by Freesurfer and SAMSEG, which supports selection of the L/R accumbens area, amygdala, caudate, hippocampus, pallidum, putamen, thalamus and ventral diencephalon. The ‘Navigate’ panel allows exploring the set of beam solutions by clicking on faces of the scalp map or connecting the tool to a Localite or Brainsight optical tracking camera. There are two main displays on the right-hand side: The top panel is a bar graph of the dose in all the sub-cortical nuclei and the lower panel is the smoothed scalp map with a 3D display of the beam at the current transducer location.

Fig. 3.

Real-time model-based tFUS navigation with optical tracking. The optical camera tracks both the subject and the transducer and streams this information into the navigation GUI, which allows real-time visualization of the tFUS beam as the operator freely moves the transducer on the subject’ scalp.

Fig. 4.

Dose deposited with line-of-sight targeting (LOST) in the left amygdala (A), right amygdala (B), left thalamus (C) and right thalamus (D) in 13 subjects using transducers with focal distances F = 55 mm, 65 mm and 80 mm. The ‘average’ bars show the average and standard deviation of deposited dose values across subjects. The min:max values indicate the range of deposited dose values across subjects for each focal distance (for example, min:max = 10:1 indicates that there is a 10-fold variation of the dose across subjects).

Fig. 5.

Dose deposited with model-based navigation (MBN) in the left amygdala (A), right amygdala (B), left thalamus (C) and right thalamus (D) in 13 subjects using transducers with focal distances F = 55 mm, 65 mm and 80 mm. The ‘average’ bars show the average and standard deviation of deposited dose values across subjects. The min:max values indicate the range of deposited dose values across subjects for each focal distance (for example, min:max = 10:1 indicates that there is a 10-fold variation of the dose across subjects).

Fig. 6.

Violin plots showing the distribution, mean and standard deviation of the acoustic dose in the left amygdala (A), right amygdala (B), left thalamus (C) and right thalamus (D) when using line-of-sight targeting (LOST), water simulation (Water) and model-based navigation (MBN) for transducers with focal distances F = 55 mm, F = 65 mm and F = 80 mm. Brackets indicate statistically significant increases of the deposited dose with MBN compared to LOST (two-sided t-test, p = 0.05).

Fig. 7.

Model-based navigation (MBN) scalp maps targeting the right thalamus with the F = 80 mm transducer. The value next to each map indicates the peak dose deposition in the target at the optimal MBN transducer position. Colormaps are scaled to that value. Optimal MBN transducer locations are shown in black, LOST locations in white.

Fig. 8.

Model-based navigation (MBN) scalp maps targeting the right amygdala with the F = 80 mm transducer. The value next to each map indicates the peak dose deposition in the target at the optimal MBN transducer position. Colormaps are scaled to that value. Optimal MBN transducer locations are shown in black, LOST locations in white.

Fig. 9.

TFUS beams targeting the right thalamus, overlaid on T1-images for 6 of the 13 subjects in this study. The first and second rows show beam solutions associated with line-of-sight targeting (LOST) and model-based navigation (MBN), respectively. The transducer is shown in dark blue (the transducer centerline ends in a small dot indicating the water focal distance, which is F = 80 mm both for LOST and MBN since this is the optimal configuration for this target as shown in Figs. 4 and 5). The outline of the target nuclei is in green. The cyan numbers indicate the dose deposited in the target region in arbitrary units.

Fig. 10.

TFUS beams targeting the right amygdala, overlaid on T1-images for 6 of the 13 subjects in this study. The first and second rows show beam solutions associated with line-of-sight targeting (LOST) and model-based navigation (MBN), respectively. The transducer is shown in dark blue (the transducer centerline ends in a small dot indicating the water focal depth, which is F = 55 mm both for LOST and F = 80 mm for MBN since these are the optimal configurations for this target as shown in Figs. 4 and 5). The outline of the target nuclei is in green. The cyan numbers indicate the dose deposited in the target region in arbitrary units.

Fig. 11.

Scalp maps of the dose in the right thalamus for 6 of the 13 test subjects, obtained with modeling of the skull (model-based navigation, MBN) and without modeling of the skull (water simulation, ‘Water’). In the ‘Water’ approach, the ideal beam profile of the transducer in water is overlapped onto the geometry of the head of the subject for all virtual transducer positions. Therefore, ‘Water’ can be viewed as a generalization of LOST. The transducer modeled has a focal distance F = 80 mm. Optimal transducer locations corresponding to the peak of those maps are shown in black.