.. _proj-desc:

.. image:: ../_static/analysis-med.jpg

*******************
Project Description
*******************


Summary
=======

We have been involved in a longitudinal research project on language
development since 2002 and it is ongoing with current support (from the
National Institute of Child Health and Human Development) through 2012. 

This project is designed to provide insights into human development and the development of language in particular.  We focus on linguistic and gestural inputs and their consequences for child language and gesture as both play out over time in typically developing and brain injured children.  Our study reveals a great deal about caregiver and child language, how each evolves over time, and how the interactions between caregivers and children influence development.

.. seealso::

    `This announcement <http://chronicle.uchicago.edu/080612/nih.shtml>`_ of
    our NIH grant provides a concise overview of our project's scope and 
    objectives.


Investigators
=============

* `Susan Goldin-Meadow
  <http://psychology.uchicago.edu/people/faculty/sgmeadow.shtml>`_, 
  *Principal Investigator*
  (`email <mailto:sgm@uchicago.edu>`__)

* `Larry Hedges
  <http://www.statistics.northwestern.edu/people/larry-hedges.html>`_ 
  (`email <mailto:l-hedges@northwestern.edu>`__)

* `Janellen Huttenlocher
  <http://psychology.uchicago.edu/people/faculty/jhuttenlocher.shtml>`_ 
  (`email <mailto:hutt@uchicago.edu>`__)

* `Susan Levine
  <http://psychology.uchicago.edu/people/faculty/slevine.shtml>`_ 
  (`email <mailto:s-levine@uchicago.edu>`__)

* `Stephen Raudenbush
  <http://home.uchicago.edu/~sraudenb/>`_ 
  (`email <mailto:sraudenb@uchicago.edu>`__)

* `Steven Small
  <http://psychology.uchicago.edu/people/faculty/ssmall.shtml>`_ 
  (`email <mailto:small@uchicago.edu>`__)


Overview and Objectives
=======================

Acquiring the ability to communicate using natural language and symbolic
gestures is a uniquely human capacity that underlies the exchange of
information among people. There is as yet no consensus concerning how
susceptible this process is to environmental and biological variation. Our
longitudinal study focuses on this issue, exploring the extent and the
limits of the language-learning process. 

To examine language growth in the face of environmental variation we have
observed 60 children, selected to represent the demographic range of 
the Chicago area, between the ages of 14 and 58 mos. and have continued to
follow them as they enter school and learn to read. Assessments have been made 
of child and parent spontaneous speech, along with narrative and reading skills from 5 to 10 years. 

Using this data, growth curves will be constructed for each child to track 
language and reading development across time, and to examine children's 
linguistic and reading progress in the later years (5-10 yrs.) in relation to 
their developmental trajectory during the early years (14-58 mos.).

To explore language growth in the face of biological variation 40 children 
with unilateral brain injury who were observed from 14 to 58 mos. are being 
followed from 5 to 10 years with an eye toward determining whether
environmental variation plays the same role in predicting their language and
reading growth as it does in children who have not suffered brain injury.

Along with traditional measures, the gesturing of our child subjects is being
examined to determine whether children who are delayed in speech relative to 
their peers use gesture to compensate for those delays.

We are also using fMRI techniques to assess the brain bases underlying linguistic and gestural competence.

Our work builds on five years of longitudinal data in a diverse sample,
and thus offers a unique opportunity to explore the impact that early
language learning has on the oral and written skills that children develop
once schooling has begun. This data has the potential to shed light on the
factors that contribute to the gap between children from high vs. low
socio-economic groups on the first day of school, and may even point to ways
of shrinking that gap.

Our major concern is with growth, and with the possibility that variations in 
caregiver speech and gesture may affect the course and outcomes of language 
development in the child. The study began in 2002 when the target children 
were 14 months old and is ongoing with current support (from the National 
Institute of Child Health and Human Development) though 2012, when all of the 
children will have completed 4th grade, and some will have completed 5th grade, 
depending on when their birthdays and time of school entry.


Participants
============

Two groups of families were enrolled at the start of this longitudinal study: 
(1) a group of 64 families with a child who was developing typically at entry
into the study (TD group) and (2) a group of 42 families with a child who had
suffered a unilateral brain lesion in the pre- or peri-natal period (PL
group).

Families of the TD children were recruited from the greater Chicago area
through a direct mailing to roughly 5,000 families or through an
advertisement in a free, monthly parent magazine.  We asked parents who
responded to participate in a phone interview in which we gathered
demographic information on their children and family.  We then selected
subjects who matched as closely as possible the ethnic/racial makeup and
family income for the Chicago area as reported in the 2000 U.S. Census.
Sixty-four families completed at least four visits, and Table 1 shows the
number of children in different income and ethnic/racial groups at entry
into the study.  In this sample, there are 31 girls and 33 boys, and 34 are
first born children.  All families participating in the study are
monolingual English speakers.

Families of children with PL were recruited through brochures in
Chicago-area pediatric neurology and rehabilitation clinics and through
advertisements to members of parent support groups for families with
children who have a brain injury.  Children were enrolled in the study if
brain imaging (MRI) results confirmed the presence of a unilateral injury of
pre- or perinatal origin. Children with congenital malformation and bleeds
associated with prematurity (under 36 weeks gestation) were excluded. On the
basis of MRI scans, lesions were categorized in terms of laterality (left,
right), size (small, medium, large) and lesion type (periventricular (PV)
lesion or cerebrovascular (CV) lesion). Small lesions affected only one lobe
or minimally affected subcortical regions.  Medium lesions extended into
more than one lobe or subcortical region.  Large lesions affected three or
four lobes, often involved the thalamus and other subcortical regions, and
were typically cerebrovascular infarcts. Regarding lesion type, CV lesions
are infarcts of the middle cerebral arty territory and tend to affect the
infereior frontal and/or superior temporal regions.  PV lesions are
primarily subcortical and involve white matter tracts, the thalamus, basal
ganglia and/or the medial temporal lobe. Because of the relative scarcity of
children with this type of injury (roughly 1 in 4,000), families were not
excluded based on demographic characteristics.  In addition, PL children
were enrolled in the study at various ages (from 14 - 54 months). Forty-two
families completed at least four visits.  In this sample, there are 25 girls
and 27 boys, 13 with right-sided brain injury, and 29 with left-sided
injury.  Different subgroups of the brain injured sample were included in
various studies depending on whether the child had participated in the study
at various ages.


Procedure 
=========

There are two parts to this study: 1) A primarily observational study of the
children and their primary caregiver(s) conducted from child age of 14 months
to 58 months. 2)  A follow-up study of the same children during the early
elementary school years (kindergarten – 4th grade) which collects data on the
child’s emerging literacy and reading skills, as well as mathematical and
spatial skills.

In the pre-school observational study, families are visited in their homes 
every four months for a total of 12 visits between 14 months and 58 months.  
The researcher videotapes the interactions of parental caregiver(s) and target 
child during their ordinary daily activities for a 90-minute period at each visit, interacting minimally with the families.  The majority of families have a
parent, usually the mother, who self-identified as the primary caregiver for
the child.  Several families (7 in TD, x in PL) are dual caregiver families, 
and these visits usually include both the mother and the father interacting 
with the target child.  Other children and family members are sometimes 
present during these visits, but our video recordings focus on the interaction 
between the target child and the parental caregiver(s).

In the school-age study, families are visited in their homes at the beginning
of every school year (Sept-Oct), the middle of the school year (Jan-Feb), and
at the end of the school year (May-June) from kindergarten through 4th grade.
At these visits, the researcher interacts one-on-one with the child over a
2-hour period, giving them a number of tasks to complete, as well as spending a
period of time (20 minutes) having a spontaneous conversation with the child.


Subject Assessments
===================

Throughout the study, children and parents were given standardized tests, tasks and questionnaires to complete.  For a description of the tasks and questionnaires given
at each timepoint, see our :doc:`Tasks page <../tasks/index>`.


Transcription of Speech and Gesture 
===================================

Transcripts are made from collected video recordings of parent/child
interaction and researcher/child interaction. Speech and gesture are 
systematically coded.   


Speech 
------

The speech utterances for both child and parent are transcribed
verbatim using English words (gotta is transcribed as got to) and incorrect
grammar is not corrected (where my puppy?). Rules were developed for
delineating utterance boundaries, including 1) an utterance is never more
than one conversational turn; 2) an utterance is never more than one
sentence long; 3) an utterance can be a single word, a phrase, or a sentence
and 4) intonational contours (such as raising the voice at the end of a
question) indicate the end of an utterance. All child speech and all primary
caregiver speech directed to the child is transcribed.  In addition,
primary caregiver speech to siblings under the age of 13 is transcribed and,
if designated as a dual caregiver, the other parent's child-directed speech
is transcribed.


Gesture 
-------

Each gesture made by the child or the parent is marked in the
transcript. Gestures are classified into 5 types (McNeill, 1992): deictic
(either a point or a hold-up), conventional (e.g., nod, side-to-side shake,
shrug), iconic (e.g., flap arms as though flying like a bird, thumb and
finger form a circle that resembles a penny), metaphoric (e.g., extending a
palm outward to represent putting an idea forward), or beats (e.g., a
rhythmic movement that punctuates speech). The form of each gesture is
described in terms of the shape of the hand, the type of movement, and the
place of articulation.  In addition, using non-linguistic context, the
first four types of gestures are assigned a meaning (see Goldin-Meadow &
Mylander 1984, for a detailed description of how meaning is assigned to
gestures). Gesture interpretation in spontaneous conversations is
facilitated by the fact that we are familiar with the activities that
typically occur during the taping sessions, and by the fact that the
parents frequently share their intimate knowledge of the child's world
with us during the taping sessions. 


Reliability
-----------

We employ two different reliability measures. The first
concerns the reliability of the transcription (inter-transcriber
agreement). For at least one out of every five transcripts, a second person 
transcribes 10% of the utterances. Agreement needs to exceed 90% at the word
level and for delineation of utterance boundaries.  The second measure concerns the intercoder agreement of particular speech or gesture categories coded. A second person codes a random selection of 10% of the utterances and the
proportion of utterances on which the two coders agreed is calculated
for each category.  Agreement needs to exceed 88% for all categories coded.