Search Abstracts | Symposia | Slide Sessions | Poster Sessions
Functional near infrared spectroscopy reveals dynamic neural recruitment across cognitive decline
Poster Session C, Thursday, October 1, 10:45 am - 12:45 pm, Wangari Maathai
Shalom Henderson1, Meryem Yucel2, Shannon Kelley2, Helen Blans1, Chae Eun Lee1, Ashley Palombizio1, Xiaoting Jiang1, Andrew Budson3, David Boas2, Swathi Kiran1; 1Center for Brain Recovery, Boston University, Boston, MA, USA, 2Neurophotonics Center, Boston University, Boston, MA, USA, 3Center for Translational Cognitive Neuroscience, Boston University, Boston, MA, USA
Introduction: Neurovascular dysfunction is among the earliest physiological changes observed in Alzheimer’s disease (AD), often preceding overt cognitive decline. Although these alterations have traditionally been studied using fMRI and PET, functional near-infrared spectroscopy (fNIRS) offers a portable, cost-effective approach for capturing cortical hemodynamics through changes in oxygenated and deoxygenated hemoglobin. Prior fNIRS studies in AD have largely focused on highly constrained tasks such as naming and demonstrated reduced task-evoked responses. However, disruptions in naturalistic speech may provide a more sensitive and ecologically valid window into early neurocognitive dysfunction. We used fNIRS to examine cortical activation during naturalistic discourse in individuals with AD and related disorders. Methods: Twenty-seven individuals with AD dementia, mild cognitive impairment (MCI), or subjective cognitive impairment (SCI), along with 10 healthy adults, participated in the study. Data were analyzed both by clinical diagnostic group and separately using MoCA-based stratification: (1) Group 1: within normal thresholds, (2) Group 2: 1 SD below, and (3) Group 3: ≥ 2 SDs below established cut-off scores. Cortical hemodynamic activity was recorded using NIRSport2 during a naturalistic discourse paradigm involving conversational speech production (i.e., experimental) and sentence repetition (i.e., control). Oxygenated hemoglobin (HbO) amplitudes were analyzed in 3-second temporal bins spanning 2 seconds pre-speech onset to 5 seconds post-speech across the left middle frontal gyrus and bilateral frontal, temporal, and parietal regions. Data were preprocessed in Homer3 using autoregressive iteratively reweighted least squares to reduce motion artifacts and serially correlated noise. Linear mixed-effects models examined group x condition x bin interactions. Results: Only statistically significant group interactions surviving false discovery rate correction (p .05) are reported. When analyzed by diagnostic group, both the SCI and dementia groups demonstrated greater left parietal activation than healthy controls at Bin 4, whereas the MCI group showed reduced right parietal activation at Bin 3. MoCA-based analyses revealed a different pattern, with the most cognitively impaired group (Group 3) exhibiting increased right frontal activation at Bins 4 and 5. Additionally, the intermediate MoCA group (Group 2) demonstrated greater condition-related modulation in the right parietal region at Bin 7 relative to cognitively intact participants (Group 1). Although greater fNIRS amplitude is generally expected during the experimental versus control condition, this three-way interaction suggests disproportionately increased right parietal recruitment in Group 2 during discourse production. Conclusion: Findings suggest distinct frontal and parietal activation patterns associated with diagnostic classification and cognitive impairment severity, including evidence of delayed or potentially compensatory recruitment in more impaired groups. Differential activation across temporal bins indicates that neural responses during discourse production are both region-specific and temporally dynamic, with greater cognitive impairment associated with delayed and greater frontal and parietal recruitment. This pattern may reflect compensatory neural engagement, whereby additional cortical resources are recruited to sustain task performance as cognitive demands increase. Collectively, these results highlight the sensitivity of fNIRS for detecting subtle alterations in functional brain organization across the cognitive aging spectrum and underscore the importance of considering both diagnostic status and cognitive severity when characterizing neural mechanisms of discourse processing.
Topic Areas: Language Production, Disorders: Acquired