Related Experiment Video
Updated: Jan 7, 2026

A Metadata Extraction Approach for Clinical Case Reports to Enable Advanced Understanding of Biomedical Concepts
Published on: September 20, 2018
Clinical Manifestations
Grace J Goodwin1, Sebastian Mehrzad2, Samantha E John3
1University of Nevada, Las Vegas, Las Vegas, NV, USA.
Background:
Behavioral variant frontotemporal dementia (bvFTD) is characterized by neuropsychiatric symptoms (NPS) and personality changes that lead to functional decline and caregiver distress. There are currently no FDA-approved pharmacotherapies for NPS in bvFTD, and existing medications, typically repurposed psychiatric and Alzheimer's agents, have limited efficacy in bvFTD. A more nuanced understanding of NPS in bvFTD is needed to inform individualized intervention. The present study uses network analysis to model NPS in bvFTD at initial clinic visit.
Method:
BvFTD patients were selected from the National Alzheimer's Coordinating Center (NACC) Uniform Data Set (UDS). Patients with other FTDs or related syndromes (e.g., corticobasal degeneration, progressive supranuclear palsy, primary progressive aphasia, amyotrophic lateral sclerosis, Alzheimer's disease) were excluded. The final sample (n = 1066) consisted of predominantly non-Hispanic white (NHW=88.6%, HW=3.2%, NH Black=2.1%, HB = .2%, other=6%,), well-educated (Med=15.2[3.14]) older adults (Mage=62.7[9.66]; 38.3% female). The NPS network consisted of 12 Neuropsychiatric Inventory Questionnaire (NPI-Q) items: delusions, hallucinations, agitation/aggression, depression/dysphoria, anxiety, elation/euphoria, apathy/indifference, disinhibition, irritability/lability, motor disturbance, nighttime behaviors, appetite/eating problems. The eLasso method was used for network estimation, where "nodes" represent binary NPI-Q items (e.g., symptom absent=0, symptom present=1) and "edges" represent their pairwise dependency. Node strength was calculated to determine relative importance of each NPI-Q item to overall network connectivity.
Result:
Most of the sample met criteria for dementia (94.5%; mild cognitive impairment=5.5%). The most frequently endorsed symptom was apathy/indifference (79.01% endorsed) followed by disinhibition (67.70%). The NPI-Q network (M = .197) consisted of positive edges, the strongest of which connected items within similar symptom domains. Four symptom clusters were identified: 1) hallucinations and delusions; 2) agitation/aggression and irritability/lability; 3) elation and disinhibition; 4) depression/dysphoria and anxiety; 5) apathy/indifference, appetite/eating problems, motor disturbance; 6) nighttime behaviors. Irritability/lability and agitation/aggression appeared to have highest node strength; however, centrality values were not stable and should be interpreted with caution.
Conclusion:
NPS in bvFTD are highly interconnected-endorsing one NPS increases the likelihood of endorsing other NPS. Interventions targeting irritability and agitation may decrease symptom connectivity and "contagion". Larger samples are needed to ensure stable centrality values and confirm network features. Findings may inform clinical presentations and management of NPS in bvFTD.
Related Concept Videos
Chronic Kidney Disease II: Clinical Manifestations
Coronary Artery Disease III: Clinical Manifestations
Endocarditis II: Clinical Features of Infective Endocarditis
Heart Failure III: Clinical Manifestations
Gastroesophageal Reflux Disease II: Clinical Features and Management
Clinical Manifestations
GERD presents itself in a multitude of ways, with symptoms varying from person to person. The hallmark symptoms are...
Hypertension III: Clinical Manifestations and Diagnostic Studies

