Los sueldos y salarios de los trabajadores del sector empresarial

Due to the nature of this study, one of its limitations is the small size of its sample. In the future, this limitation may be overcome with the incorporation of similar datasets that are currently under construction, such as the Carolinas’ Conversations Collection in English (Pope & Davis, 2011) and Spanish (Hernández Domínguez, Ratté, Pope, & Davis, 2016), and the CorpAGEst (Bolly & Boutet, 2018). This inclusion would not only augment the number of participants, but also allow for deeper inter-language comparisons.

Another limitation of this work is that since it was originally intended to describe the language evolution in healthy older individuals, there was no medical history or formal cognitive assessments of the participants available. The diagnoses of the cognitive impairment group were annotated as described by the spouses, family members and caregivers of the participants. Still, the nature of the speech data seems to balance these limitations, as the individuals did not perceive themselves as patients; the stereotypes underlying cognitive testing and the supposed decline in old age could lead to a lower language production performance (Hess, Hinson, & Hodges, 2009).

4.7 Conclusions

In this work, we presented a computer-based longitudinal analysis of linguistic changes in older French speakers with cognitive impairment. This study was done using a sample from the LangAge corpus (Gerstenberg, 2011), in which three sets of interviews were conducted with the participants over a span of ten years. All participants where apparently healthy individuals at the time of the first interview, and four developed cognitive impairment by the end of the process. Through a computer-based quantitative analysis, we compared the evolution of language alterations in these four participants against four healthy control participants matched by age, gender, educational level and bilingualism. To the best of our knowledge, this is the first work using spontaneous conversations to study the language changes occurring in participants as they transition from a healthy, pre-clinical stage, to cognitive impairment.

Few previous works have been conducted on spontaneous conversations, with those done mostly being for English speakers. In this study, we extracted the main characteristics that have consistently been found to be altered by the presence of CI and studied their evolution over time. We also proposed the use of four linguistic measures based on vocabulary distribution to observe the use of general and specific vocabulary by the speakers.

We found that most participants in the CI impaired group presented a continuous drop in the lexical richness measures throughout the ten years of the study. Also, for most metrics, HC individuals started with better scores beginning with the first interview. Our proposed metrics based on vocabulary distribution showed a tendency of CI participant to reduce the number of original words and phrases over time, and a predisposition to talk about similar topics through most interviews. It also showed that CI participants will tend to use more generic vocabulary and use more pronouns and fewer nouns and verbs, except for common use verbs.

After a PCA, we found that two main components were highly correlated with the severity of cognitive impairment. In a scatter plot we observed a clear separation of the severely impaired individuals from the rest. Furthermore, with the exception of one participant, there also seems to be a clear separation between apparently healthy individuals that later developed CI from those who remained healthy throughout the ten-year span.

One apparently healthy participant had a pattern of decline in most metrics similar to that in the CI group. Despite this tendency, this participant consistently had higher scores in the metrics than his CI counterpart. This participant is a highly educated individual who held a high academic or management profession, factors that have been linked to a higher cognitive reserve (Almeida et al., 2015). In 2015, he reportedly underwent a medical exam, performed by a neurologist, confirming the absence of CI indicators. At the time of writing of this paper, there was no evidence that the participant had developed any form of CI; however, it would be beneficial to monitor this individual’s progression.

It is our belief that our proposed method and analysis could help in the following up of patients trough time and develop more personalized analyses of cognitive status. The study of spontaneous speech represents an inexpensive and non-invasive process for detecting early signs of cognitive impairment. Our metrics evinced a significant difference between individuals that would age as cognitively intact individuals and those who would develop a form of cognitive impairment even up to ten years before the time of diagnosis.

CONVERSING WITH THE ELDERLY IN LATIN AMERICA: A NEW COHORT FOR MULTIMODAL, MULTILINGUAL LONGITUDINAL STUDIES ON AGING

Laura Hernández-Domínguez1_{, Sylvie Ratté}1_{, Charlene Pope}2 _{and Boyd Davis}3

1 _{Software and IT Engineering Department,}

École de technologie supérieure, Montreal, Canada

2 _{Medical University of South Carolina (MUSC), Charleston, SC, United States of America} 3 _{University of North Carolina at Charlotte, Charlotte, NC, United States of America}

Email: [email protected], [email protected]

This article was published at the ACL’s 7th Workshop on Cognitive Aspects of

Computational Language Learning, in Berlin, Germany, on August 11, 2016.

Foreword: As observed I the literature review and in the limitations presented in Chapter 4, one of the biggest obstacles for performing studies on longitudinal analyses of language alterations in spontaneous conversations is the scarcity of data available for research. Besides the direct study of these alterations, a concomitant contribution of my doctoral studies was oriented to increase the available data of this nature by collaborating in the creation of the Latin-American cohort of the Carolinas’ Conversation Collection. The recollections for this cohort started in 2015, and although this data was not used for analysis, since the recollections and transcriptions are still in process, it is expected that it will be used for the continuation of the multi-modal analyses for the Cécilia Project.

5.1 Abstract

Many studies have found that language alterations can aid in the detection of certain medical afflictions. In this work, we present an ongoing project for recollecting multilingual conversations with the elderly in Latin America. This project, so far, involves the combined efforts of psychogeriatricians, linguists, computer scientists, research nurses and geriatric

caregivers from six institutions across USA, Canada, Mexico and Ecuador. The recollections are being made available to the international research community. They consist of conversations with adults aged sixty and over, with different nationalities and socio-economic backgrounds. Conversations are recorded on video, transcribed and time-aligned. Additionally, we are in the process of receiving written texts---recent or old---authored by the participants, provided voluntarily. Each participant is recorded at least twice a year to allow longitudinal studies. Furthermore, information such as medical history, educational background, economic level, occupation, medications and treatments is being registered to aid conducting research on treatment progress and pharmacological effects. Potential studies derived from this work include speech, voice, writing, discourse, and facial and corporal expression analysis. We believe that our recollections incorporate complementary data that can aid researchers in further understanding the progression of cognitive degenerative diseases of the elderly.

5.2 Introduction

The Carolinas Conversations Collection (Pope & Davis, 2011), a project for recollecting conversations with elderly people that live in North and South Carolina, started in 2008. This project was initially supported by the USA National Library of Medicine. For the collection, the conversations were transcribed, marked, time-aligned and made available to the international research community by means of a secured website3_{. The collection has grown}

steadily since then, having, at present, over 460 conversations with adults over sixty years old, either healthy or suffering from any medical condition. A fourth of these conversations were made with participants afflicted with Alzheimer's disease.

In 2015, we started to increase the coverage of this collection to incorporate different languages. The first additional language to be incorporated is Latin-American Spanish. We are currently adding conversations with new participants; elderly Spanish speakers from Ecuador and Mexico. Additionally, we are incorporating new information and language modalities to

increase the robustness of possible studies that may use this corpus. So far, this project has engaged involvement through combined efforts of six institutions across four different countries.