Poster | Poster Session 06 Program Schedule
02/15/2024
04:00 pm - 05:15 pm
Room: Majestic Complex (Posters 61-120)
Poster Session 06: Aging | MCI | Neurodegenerative Disease - PART 2
Final Abstract #65
Semantic Network Interconnectivity Impacts Word Retrieval in Multiple Sclerosis
Sophia Lall, MA, Montclair State University, Montclair, United States Joshua Sandry, PhD, Montclair State University, Montclair, United States
Category: Multiple Sclerosis/ALS/Demyelinating Disorders
Keyword 1: multiple sclerosis
Keyword 2: semantic processing
Keyword 3: language
Objective:
Word finding difficulty is a common concern among persons with Multiple Sclerosis (pwMS). However, assessment of semantic verbal fluency yields mixed results in the detection of word retrieval difficulties. Additional information concerning word retrieval processes can be obtained through clustering and switching, two strategies that respectively represent memory/executive function and attention/executive function abilities. Prior research, although mixed, largely supports decreased switching in pwMS.
Semantic network analysis may provide additional insight into word retrieval difficulties, as it utilizes the co-occurrence data of words produced. Resulting networks provide an estimation of the organization of the conceptual-lexical store. A preliminary study found that semantic networks from pwMS had fewer connections and were more structurally vulnerable. Although participants were matched on demographic factors, results were limited as authors did not equate the two groups’ networks on their number of nodes and edges. A recently published study equated the group networks and found that pwMS exhibited higher average shortest path length (ASPL), lower clustering coefficient (CC), and higher modularity (Q) relative to controls. These features suggest reduced efficiency and interconnectivity, as well as increased organization into subcommunities. An important limitation was that the MS sample had significantly lower educational attainment and greater ethnic diversity than controls. As education and ethnicity are significant predictors of task performance, it is unclear whether results reflect true group differences or the impact of these factors. We sought to replicate the results from this recent study.
Participants and Methods:
75 pwMS and 65 neurologically healthy controls (HC) were recruited as a part of an ongoing study (NMSS RG-1907-34364 & RG-1901-33304). As a part of a larger battery, participants were administered the Animals verbal fluency test. Analyses were carried out using R and Python.
Results:
Groups were not significantly different with regard to age (t(104.85)=-1.28, p=0.20), sex (χ2(1, N=140)=2.83, p=0.09), education (t(138)=0.74, p=0.46), and ethnicity (Fisher’s exact test; p=0.40). Total number of responses were significantly lower in the MS group (MS=21.77, HC=23.98, t(138)=2.14, p=0.03). Average cluster size (MS=1.93, HC=1.95) and number of cluster switches (MS=11.00, HC=12.03) were similar between the MS and HC groups. Consistent with prior findings, the MS network was characterized by having a significantly lower clustering coefficient (CCMS=0.747, CCHC=0.749, F(1,1997)=39.196, p<0.001). In contrast to prior findings, both MS and HC networks displayed similar average shortest path length (ASPLMS =2.542, ASPLHC =2.536, F(1,1997)=0.842, p=0.359) and modularity (QMS=0.518, QHC=0.518, F(1,1997)=0.393, p=0.531).
Conclusions:
Results support the presence of semantic word retrieval difficulty in pwMS. Fewer words were retrieved, which may be driven by reduced interconnectivity in the MS network. Null findings concerning clustering and switching were consistent with network findings of similar modularity between groups. This implies that both groups were equally able to utilize categories and switch between categories to retrieve words during the task. The prior study reported ASPL was higher in pwMS; our results showed the same directionality but were not significant. Modularity was higher in pwMS in the prior study; our results showed no difference. It is possible that differing results across studies are due, at least in part, to sample demographic differences.
|