Portal dosimetry of small unflattened beams.

We developed and validated a dedicated small field back-projection portal dosimetry model for pretreatment andin vivoverification of stereotactic plans entailing small unflattened photon beams. For this purpose an aSi-EPID was commissioned as a small field dosimeter. Small field output factors for 6 MV FFF beams were measured using the PTW microDiamond detector and the Agility 160-leaf MLC from Elekta. The back-projection algorithm developed in our department was modified to better model the small field physics. The feasibility of small field portal dosimetry was validated via absolute point dose differences w.r.t. small static beams, and 5 hypofractionated stereotactic VMAT clinical plans measured with the OCTAVIUS 1000 SRS array dosimeter and computed with the treatment planning system Pinnacle v16.2. Dose reconstructions using the currently clinically applied back-projection model were also computed for comparison. We found that the latter yields underdosage of about -8% for square beams with cross section near 10 mm x 10 mm and about -6% for VMAT treatments with PTV volumes smaller than about 2cm3. With the methods described in this work such errors can be reduced to less than the ±3.0% recommendations for clinical use. Our results indicate that aSi-EPIDs can be used as accurate small field radiation dosimeters, offering advantages over point dose detectors, the correct positioning and orientation of which is challenging for routine clinical QA.

Uncertainty Propagation in Nerve Impulses Through the Action Potential Mechanism.

We investigate the propagation of probabilistic uncertainty through the action potential mechanism in nerve cells. Using the Hodgkin-Huxley (H-H) model and Stochastic Collocation on Sparse Grids, we obtain an accurate probabilistic interpretation of the deterministic dynamics of the transmembrane potential and gating variables. Using Sobol indices, out of the 11 uncertain parameters in the H-H model, we unravel two main uncertainty sources, which account for more than 90 % of the fluctuations in neuronal responses, and have a direct biophysical interpretation. We discuss how this interesting feature of the H-H model allows one to reduce greatly the probabilistic degrees of freedom in uncertainty quantification analyses, saving CPU time in numerical simulations and opening possibilities for probabilistic generalisation of other deterministic models of great importance in physiology and mathematical neuroscience.

Error probability of intracranial brain computer interfaces under non-task elicited brain states.

OBJECTIVE: Intracranial brain computer interfaces (BCIs) can be connected to the user's cortex permanently. The interfaces response when fed with non-task elicited brain activity becomes important as design criterion: ideally intracranial BCIs should remain silent. We study their error probability in the form of false alarms. METHODS: Using electrocorticograms recorded during task and non-task brain states, we compute false alarms, investigate their origin and introduce strategies to reduce them, using signal detection theory, classifier cascading and adaptation concepts. RESULTS: We show that the incessant dynamics of the brain is prone to spontaneously produce signals, the spectral and topographical characteristics of which can resemble those associated with common control tasks, generating brain state classification errors. CONCLUSIONS: In addition to hit and bit rates, response of BCIs to non-task brain states constitutes an important measure of BCI performance. Static classification cascading reduces considerably false positives during no-task brain states. SIGNIFICANCE: False alarms in intracranial BCIs are undesirable and could have dangerous consequences for the users. Designs which effectively incorporate the error correction strategies discussed in this paper, could be more successful when taken from the laboratory or acute care setting and used in the real world.

Gain of the human dura in vivo and its effects on invasive brain signal feature detection.

Invasive brain signal recordings generally rely on bioelectrodes implanted on the cortex underneath the dura. Subdural recordings have strong advantages in terms of bandwidth, spatial resolution and signal quality. However, subdural electrodes also have the drawback of compromising the long-term stability of such implants and heighten the risk of infection. Epidurally implanted electrodes might provide a viable alternative to subdural electrodes, offering a compromise between signal quality and invasiveness. Determining the feasibility of epidural electrode implantation for e.g., clinical research, brain-computer interfacing (BCI) and cognitive experiments, requires the characterization of the electrical properties of the dura, and its effect on signal feature detection. In this paper we report measurements of brain signal attenuation by the human dura in vivo. In addition, we use signal detection theory to study how the presence of the dura between the sources and the recording electrodes affects signal power features in motor BCI experiments. For noise levels typical of clinical brain signal recording equipment, we observed no detrimental effects on signal feature detection due to the dura. Subdural recordings were found to be more robust with respect to increased instrumentation noise level as compared to their epidural counterpart nonetheless. Our findings suggest that epidural electrode implantation is a viable alternative to subdural implants from the feature detection viewpoint.