Biology subjects

Archila-Melendez, M. E.

Publications and source records attributed to Archila-Melendez, M. E..

3 recordsLinked to original sources

bioRxiv2026

Item-specific clustering of high gamma-band activity on theta phase during speech categorization

Theta-gamma phase-amplitude coupling (PAC) has been proposed as a mechanism for separating neural representations in time. Although item-specific PAC has been reported, its role in perception remains unclear. Notably, previous work has shown that theta phase can bias perception, raising the question whether such biases are linked to phase-specific gamma activity. Here, we investigated whether phase-dependent perceptual biases are accompanied by item-specific theta-gamma PAC. We presented ambiguous speech sounds that could be interpreted as either /da/ or /ga/ in patients with epilepsy implanted with intracranial electrocorticography. We found that pre-stimulus theta phase predicted both the neural and perceptual response, with /da/ and /ga/ tuned channels exhibiting gamma activity clustering at distinct theta phases. These findings suggest that neural and perceptual biases stem from phase-specific pre-activation of item-specific neural representations. They demonstrate the generalizability of the theta-gamma code, shaping behavior and enhancing representational capacity.

neuroscience↗

bioRxiv2025

In vivo reconstruction of Duvernoy's postmortem vasculature images

Non-invasive measurement of the human brains angioarchitecture is critical for understanding the basis of functional neuroimaging signals, diagnosing cerebrovascular diseases, and tracking neurodegeneration. Ultra-high-field magnetic resonance imaging (MRI) has achieved mesoscopic (< 0.5 mm) imaging of angioarchitecture, revealing fine vascular details that were previously inaccessible in vivo. However, current mesoscopic MRI methods for imaging angioarchitecture face two major limitations. First, acquisition times are prohibitively long, often exceeding 40 minutes, making integration into everyday clinical practice and research projects impractical. Second, even with data successfully acquired, conventional data visualization methods, such as 2D slice browsing and 3D vessel segmentation renders, are rudimentary and have limited effectiveness for navigating and interpreting the complex vascular network. In this paper, we present a fast whole-brain MRI protocol that provides robust images of the brains venous network at 0.35 mm resolution in under seven minutes. Additionally, we introduce novel data processing and visualization techniques that enable identification of specific vessel types and more informative navigation of the complex vascular network. We demonstrate that, with these advancements, we can reproduce, in vivo and without intravenous contrast application, the seminal postmortem vasculature images of Duvernoy and Vannson (1999). Furthermore, leveraging the ability for MRI to cover the entire brain, we achieve, for the first time, whole-brain intracortical mesoscopic vein maps in humans. Our acquisition and post-processing methods lay the groundwork for detailed examination of vascular organization across individuals, brain regions, and cortical layers. More generally, these methods make mesoscopic imaging of angioarchitecture viable for broad neuroscientific and clinical applications.

neuroscience↗

bioRxiv2019

A novel approach for combining task-dependent gamma with alpha and beta power modulation for enhanced identification of eloquent cortical areas using ECoG in patients with medical-refractory epilepsy

Electrical stimulation mapping (ESM) is the gold standard for identification of \"eloquent\" areas prior to resection of epileptogenic tissue, however, it is time consuming and may cause side effects, especially stimulation-induced seizures and after-discharges. Broadband gamma activity (55 - 200 Hz) recorded with subdural electrocorticography (ECoG) during cognitive tasks has been proposed as an attractive tool for mapping cortical areas with specific function but until now has not proven definitive clinical value. Fewer studies have addressed whether the alpha (8 - 12 Hz) and beta (15 - 25 Hz) band activity could also be used to improve eloquent cortex identification. We compared alpha, beta and broadband gamma activity, and their combination for the identification of eloquent cortical areas defined by ESM. Ten patients participated in a delayed-match-to-sample task, where syllable sounds were matched to visually presented letters and responses given by keyboard. We used a generalized linear model (GLM) approach to find the optimal weighting of low frequency bands and broadband gamma power to predict the ESM categories. Broadband gamma activity increased more in eloquent areas than in non-eloquent areas and this difference had a diagnostic ability (area under (AU) the receiving operating characteristic curve - AUROC) of [~]70%. Both alpha and beta power decreased more in eloquent areas. Alpha power had lower AUROC than broadband gamma while beta had similar AUROC. AUROC was enhanced by the combination of alpha and broadband gamma (3% improvement) and by the combination of beta and broadband gamma (7% improvement) over the use of broadband gamma alone. Further analysis showed that the relative performance of broadband gamma and low frequency bands depended on multiple factors including the time period of the cognitive task, the location of the electrodes and the patients attention to the stimulus. However, the combination of beta band and broadband gamma always gave the best performance. We show how ECoG power modulation from cognitive testing periods can be used to map the probability of eloquence by ESM and how this probability can be used as an aid for optimal ESM planning. We conclude that low frequency power during cognitive testing can contribute to the identification of eloquent areas in patients with focal refractory epilepsy improving its precision but does not replace the need of ESM.\n\nHighlightsO_LIGamma, alpha and beta band activity has significant diagnostic ability to identify ESM defined eloquent cortical areas.\nC_LIO_LIWe present a novel method to combine gamma and low frequency activity for enhanced identification.\nC_LIO_LIWe quantify how identification is dependent on analysis time window, cortical function, and patients attentional engagement.\nC_LIO_LIWith further development, this approach may offer an alternative to ESM mapping with reduced burden for patients.\nC_LI

neuroscience↗