Fig 1: Multimodal characterization of human cortical access tissue from preoperative MRI and Raman-based imaging to functional and ultrastructural analyses. (A) Cortical access tissue from resection of a right temporal IDH-mutant astrocytoma (WHO grade 2). Tumor marked by high intensity in the FLAIR signal. Lower left panel shows a 250 μm tissue slice undergoing stimulated Raman histology (inset). Right panel: pseudo H&E-like stimulated Raman histology (SRH). Imaging was performed using NIO microscope (Invenio Imaging, Santa Clara, USA). Visualization of nuclei, myelinated axons, and intracellular lipid-rich organelles. (B) Tumor-free neocortical access tissue obtained during surgical resection of an intraventricular glioblastoma (WHO grade 4). Arrowheads represent the location of cortical access tissue marked by neuronavigation (Stryker, Kalamazoo, USA). (C) Tumor-infiltrated access tissue derived from a non-contrast enhancing IDH-mutant astrocytoma (WHO grade 2). (D) Volumetric SRS imaging acquired in the CH2 stretching band, predominantly representing lipid-rich structures (myelinated axons). Cortical tissue originated from the parieto-occipital lobe and was resected during surgical treatment of pharmacoresistant epilepsy (focal cortical dysplasia type IIA). Imaging was performed using a Leica Stellaris CRS system (Leica, Wetzlar, Germany). (E) Experimental setup for repetitive (transcranial) magnetic stimulation and electrophysiology. (F) Exemplary miniature excitatory postsynaptic current (mEPSC) whole-cell patch clamp recording from which event frequencies and amplitudes are extracted. Representative confocal laser scanning microscopy image showing interaction of microglia (aIba1-green) and astrocytes (aGFAP-magenta) identified by immunohistochemical labeling. (G) TEM bright-field imaging provides high-resolution 2D ultrastructural information but is limited in axial (z) resolution due to section thickness (~60 nm, ZEISS LEO 906E). Synaptic features, including presynaptic boutons, free and docked synaptic vesicles, postsynaptic densities (PSD), and postsynaptic spines, can be reliably identified and quantified. Expert-labeled datasets derived from TEM and FIB-SEM images are used to train and refine automated segmentation pipelines based on deep learning approaches, for example, synaptic vesicle detection using YOLO-based object detection (Diwan et al., 2023). (H) FIB-SEM tomography provides (near-)isotropic voxel resolution with superior z-resolution compared to TEM and is used to generate accurate 3D reconstructions of synaptic and subcellular structures (Thermo Scientific Helios 5 CX). This complements TEM by enabling targeted volumetric imaging of selected regions within multimodal experimental workflows. A small volume of human cortical tissue and a few segmented neurons are shown (Fedorov et al., 2012).