Toxoplasmosis Beyond Transplantation: Diagnostic and Prevention Challenges in a Patient Receiving Targeted Immunomodulators.

Despite advances in imaging technology, continuously capturing and interpreting wide-field-of-view (FOV) visual information in complex environments remains a fundamental challenge. This is particularly true under low-light conditions, where weak object signals hinder reliable detection and tracking, limiting applications in mixed reality, visual restoration, and autonomous machine vision. Over millions of years of evolution, fruit flies () have evolved a visual system that integrates optics, motor control, and neural processing. Unlike single-chamber human eyes, which rely on plane-to-plane projection, each fruit fly possesses two natural compound eyes (NCEs), each covering an FOV of up to 180° (Fig. ). NCEs are composed of densely packed ommatidia-the basic optical units (Fig. ) arranged on a curved surface, enabling image formation primarily through angular projection (Fig. ) . Beyond static optics, NCEs are assisted by motor control: in dim environments, head bending driven by muscles extends the FOV, enhances light capture, and reduces motion blur, thereby allowing continuous tracking of moving objects (Supplementary Movie ). Furthermore, recent studies indicate that visual signals from NCEs are processed by neurons to reconstruct corresponding virtual scenes in the brain. Together, the angular projection optics, the motor control, and the neural processing of NCEs provide a robust biological blueprint for developing next-generation flexible artificial compound eyes (ACEs). However, most existing ACE designs have primarily focused on mimicking the function of NCEs, largely neglecting the critical roles of motor control and neural processing, which are so-called typical ACEs. The head of features a pair of Natural compound eyes (NCEs) for light reception, muscles for head bending, and neural circuits for visual processing. Each NCE is composed of numerous natural ommatidia, surrounded by pigment cells that prevent optical crosstalk. The interommatidial angle ∆ = /, where and denote the arc distance of adjacent ommatidia and the local radius of curvature, respectively. Within a natural ommatidium, a facet lens with focal length collects light within a specific acceptance angle ∆. The crystalline cone focuses the light onto the rhabdom (diameter ), which transmits it through the inner structure, while the photoreceptor cell detects and records the incoming light. The FACEcam is mounted on a mixed reality (MR) device to enable wide-angle light perception and signal processing. Two tethers are used to bend the head of the FACEcam, mimicking the head-bending muscles of . SCL: shallow convolution layer mimicking lamina; DCL: deep convolution layer mimicking medulla; FCL: fully connected layer mimicking lobula. The artificial compound eye is composed of numerous artificial ommatidia. A flat imaging sensor chip simulates the deep neural centres, where the captured signals are processed using artificial intelligence algorithms. Each artificial ommatidium closely replicates the function of a natural one: a microlens mimics the facet lens to restrict the acceptance angle, defined as ∆ = /, where is the optical fibre core diameter and is the focal length of the microlens. An optical fibre core imitates both the crystalline cone and the rhabdom to transmit light, while the cladding mimics pigment cells to block optical crosstalk. An imaging lens focuses light from each fibre onto a specific photodetector, and the photodetector on the flat sensor mimics the photoreceptor cell to record visual information. A structural flange ensures that the distal ends of the optical fibres sit on the curved dome surface. The evolutionary path of ACEs in the functional mimicry of NCEs, head muscles, and neurons. Typical ACEs can be traced back to the planar microlens arrays, whose FOV is restricted due to their flat configuration (Fig. ), and curved microlens arrays (Fig. ), which can be further classified into ACEs employing planar photodetector arrays (PDAs) and curved PDAs, mimicking the distribution of microlenses in the outermost layer of NCEs (Fig. ). Although ACEs with planar PDAs extend the FOV to approximately 140° by leveraging the curved distribution of microlenses , they still suffer from significant off-axis aberrations. To address this, researchers have attempted to introduce curved PDAs, enabling light from individual microlenses to be more accurately captured, thereby extending the FOV to around 160° and minimising off-axis aberrations . Nonetheless, these designs struggle to achieve both real-time imaging and dynamic motion detection simultaneously due to the discrete nature of the photodetectors. Further improvements are seen in ACEs integrating light guides (Fig. ), which facilitate light transmission between microlenses and PDAs, achieving an FOV of up to 180°, and faithfully mimicking the anatomic structure and function of NCEs (Fig. ). Despite these advancements, the FOV remains confined to a hemisph