Automated morphological analysis of clinical language samples

Kyle Gorman,Metrah Mohammed,Steven Bedrick,Jan Van Santen,Rosemary Ingham,Eric Morley,Katina Papadakis,Geza Kiss

doi:10.3115/v1/w15-1213

Abstract

Quantitative analysis of clinical language samples is a powerful tool for assessing and screening developmental language impairments, but requires extensive manual transcription, annotation, and calculation, resulting in error-prone results and clinical underutilization. We describe a system that performs automated morphological analysis needed to calculate statistics such as the mean length of utterance in morphemes (MLUM), so that these statistics can be computed directly from orthographic transcripts. Estimates of MLUM computed by this system are closely comparable to those produced by manual annotation. Our system can be used in conjunction with other automated annotation techniques, such as maze detection. This work represents an important first step towards increased automation of language sample analysis, and towards attendant benefits of automation, including clinical greater utilization and reduced variability in care delivery.

Highlights

Specific language impairment (SLI) is a neurodevelopmental disorder characterized by language delays or deficits in the absence of other developmental or sensory impairments (Tomblin, 2011)
There has been a recent push to augment norm-referenced tests with language sample analysis (Leadholm and Miller, 1992; Miller and Chapman, 1985), in which a spontaneous language sample collected from a child is used to compute various statistics measuring expressive language abilities
Our evaluation demonstrates that this model produces estimates of mean length of utterance in morphemes (MLUM) which are very similar to those produced by manual annotation

Summary

Introduction

Specific language impairment (SLI) is a neurodevelopmental disorder characterized by language delays or deficits in the absence of other developmental or sensory impairments (Tomblin, 2011). Developmental language impairments are normally assessed using standardized tests such as the Clinical Evaluation of Language Fundamentals (CELF), a battery of norm-referenced language tasks such as Recalling Sentences, in which the child repeats a sentence, and Sentence Structure, in which the child points to a picture matching a sentence. There has been a recent push to augment norm-referenced tests with language sample analysis (Leadholm and Miller, 1992; Miller and Chapman, 1985), in which a spontaneous language sample collected from a child is used to compute various statistics measuring expressive language abilities. Some recent work has applied NLP techniques to quantify clinical impressions that once were merely qualitative (e.g., Rouhizadeh et al 2013, van Santen et al 2013) and other work has proposed novel computational features for detecting language disorders (e.g., Gabani et al 2011).

Objectives

Results

Discussion

Conclusion