Journal: bioRxiv

Article Title: High speed functional imaging with a microfluidics-compatible open-top light-sheet microscope enabled by model predictive control of a tunable lens

doi: 10.1101/2025.07.23.666439

Figure Lengend Snippet: Open-top light-sheet microscope with electrically tunable lens (ETL) remote focusing under model predictive control. (a) Scan methods for volumetric light-sheet imaging. (i) In the simplest case, a sample is scanned through a static sheet. (ii) An actuated objective lens can follow a scanning light-sheet, which is faster than (i) but inertia limited and can cause problems with water immersion. (iii) Remote focusing optically shifts the focal plane using an active element like an ETL (b). This setup has less inertia and no moving parts at the sample. (c) Long-working distance open-top imaging is achieved using an asymmetric pair of objective lenses coupled with a water immersion fitting. This enables unobstructed imaging across a water-matched barrier such as FEP (above). (d) Maximum intensity projections along the Z (top) and Y (bottom) axes of a worm expressing a pan-neuronal nuclear-localized fluorescent protein positioned in a microfluidic channel as shown in (c). Scale bar 20 µ m. (e) Fast actuation of an ETL induces high frequency oscillation which slows response time. Performance is improved by using model predictive control to optimize drive signals. The controller iteratively optimizes the input signal to the ETL by minimizing simulated output error while obeying system constraints.

Article Snippet: Finally, we used the generated model to construct a model predictive controller using the MATLAB Model Predictive Control toolkit.

Techniques: Microscopy, Control, Imaging, Expressing

Journal: Nature Communications

Article Title: Persistent activity in human parietal cortex mediates perceptual choice repetition bias

doi: 10.1038/s41467-022-33237-5

Figure Lengend Snippet: Effect sizes from a general linear mixed model (see “Methods”) estimating impact of stimulus category ( a ), choice ( b – d ), and action ( e – h ) on MEG power, in different regions and frequency bands. a Effect of current stimulus category on gamma (top) and alpha (bottom) power during test stimulus ( n = 60). b As a , but for effect of previous choice in the reference interval. c As b , but for test stimulus interval. d Comparison of choice history signals between alternators ( n = 25; orange) and repeaters ( n = 34; purple). Group effect = 0.55328, CI [0.19342, 0.91314], p = 0.00258. e Effect of current action preparation on gamma (top) and beta (bottom) power during test stimulus ( n = 60). f As e , but for effect of previous action during reference interval. g As f , but for the test interval. h Comparison of action history signals between alternators ( n = 25; orange) and repeaters ( n = 34; purple). Data are shown as average fixed effect size +/− 95% confidence intervals; *0.01 < p < 0.05 (no such effect present); **0.001 < p < 0.01; *** p < 0.001; filled markers, p < 0.05 (all p values FDR corrected).

Article Snippet: We used general linear mixed effects models (GLMEs, using Matlab’s fitglme ) to quantify the effect of choice history on single-trial power modulation values across all source-level ROIs, frequency ranges and the above-defined time windows.

Techniques: Comparison

Journal: Nature Communications

Article Title: Persistent activity in human parietal cortex mediates perceptual choice repetition bias

doi: 10.1038/s41467-022-33237-5

Figure Lengend Snippet: Time-resolved effect sizes from a general linear mixed model (see “Methods”), separately for both subgroups. a Effect of current and previous choice on IPS2/3 gamma-band activity. b Effect of current and previous choice on IPS0/1 alpha-band activity. c Effect of current and previous action on motor lateralization (pooled signal from M1, PMd/v, IPS/PCeS) in the beta-band. Lower markers indicate timepoints where the fixed effect is significantly different within each group, or between groups ( p < 0.05, FDR-corrected). Data are shown as average fixed effect size +/− 95% confidence intervals.

Article Snippet: We used general linear mixed effects models (GLMEs, using Matlab’s fitglme ) to quantify the effect of choice history on single-trial power modulation values across all source-level ROIs, frequency ranges and the above-defined time windows.

Techniques: Activity Assay