Measuring Word Recognition Performance Using an Innovative Computer Algorithm to Maximize Clinical Efficiency
Location
D.P. Culp Center Room 304
Start Date
4-5-2024 1:30 PM
End Date
4-5-2024 2:30 PM
Name of Project's Faculty Sponsor
Marc Fagelson
Faculty Sponsor's Department
Audiology and Speech Language Pathology
Competition Type
Competitive
Type
Oral Presentation
Presentation Category
Health
Abstract or Artist's Statement
This study focused on the word recognition test, which is a vital part of the audiologic test battery. This task asked participants to repeat monosyllabic words as a test of the patient’s ability to recognize words in a quiet background. While this portion of an audiologic exam contains valuable information, it is time consuming and is on occasion shortened to optimize appointment time. A computer algorithm to help improve this testing process is in development, but in order to make it as effective as possible, certain aspects of the word recognition protocol need to be understood in more detail. This study specifically looked at 3 aspects of this test. The first was that once a word was recognized correctly at a lower stimulus level, at higher levels it would continue to be correctly identified. The second aspect was to study how higher presentation levels would influence the overall word recognition score. The final aspect related to the number of words the subject remembered after the task was completed. Each participant was presented one fifty-word NU-6 list per ear, and each list was presented six times at levels ranging from 0 dB HL to 36 dB HL. A computerized program was utilized to administer this task. The study found that as presentation level increased, the overall recognition score increased until it plateaued. The total number of words remembered ranged from three to forty, with the average number of words remembered being 17. On average, if a word was recognized at a low presentation level, it was recognized at higher presentation levels 86.75% of the time. There was no significant correlation between performance on the overall word recognition test and total words remembered. This data helped confirm some assumptions about the word recognition test that have never been assessed and should improve the word recognition task in the future, making it more efficient and accurate.
Measuring Word Recognition Performance Using an Innovative Computer Algorithm to Maximize Clinical Efficiency
D.P. Culp Center Room 304
This study focused on the word recognition test, which is a vital part of the audiologic test battery. This task asked participants to repeat monosyllabic words as a test of the patient’s ability to recognize words in a quiet background. While this portion of an audiologic exam contains valuable information, it is time consuming and is on occasion shortened to optimize appointment time. A computer algorithm to help improve this testing process is in development, but in order to make it as effective as possible, certain aspects of the word recognition protocol need to be understood in more detail. This study specifically looked at 3 aspects of this test. The first was that once a word was recognized correctly at a lower stimulus level, at higher levels it would continue to be correctly identified. The second aspect was to study how higher presentation levels would influence the overall word recognition score. The final aspect related to the number of words the subject remembered after the task was completed. Each participant was presented one fifty-word NU-6 list per ear, and each list was presented six times at levels ranging from 0 dB HL to 36 dB HL. A computerized program was utilized to administer this task. The study found that as presentation level increased, the overall recognition score increased until it plateaued. The total number of words remembered ranged from three to forty, with the average number of words remembered being 17. On average, if a word was recognized at a low presentation level, it was recognized at higher presentation levels 86.75% of the time. There was no significant correlation between performance on the overall word recognition test and total words remembered. This data helped confirm some assumptions about the word recognition test that have never been assessed and should improve the word recognition task in the future, making it more efficient and accurate.