Computer Analysis of Speech
for Psychological Research


Michael A. Covington, Principal Investigator
Institute for Artificial Intelligence
The University of Georgia





Language is a window into the brain.

The goal of the CASPR project is to find ways to
diagnose, assess, and investigate mental illness
and other brain impairments

through computer analysis of speech.

CASPR's initial work was funded by GlaxoSmithKline Plc.

Use the following links to learn more about it.


Software downloads

NLPUtils (Boisclair tokenizer and morphological analyzer)
Verstaile tokenizer and morphological analyzer for English, in C#.
Programs and documentation (updated 2008 March 21)
Project web site

MATTR (Moving-Average Type-Token Ratio) version 2.0 for Windows
(functionally like 1.0, slightly faster)
Documentation (PDF, Research Report 2007-03, unchanged since MATTR 1.0)
Executable and source code for Windows 2000/XP/Vista
(believed to be usable under Mac OS X and Linux by same methods as CPIDR 3; see manual)

MATTR (Moving-Average Type-Token Ratio) version 1.0 for Windows
Documentation (PDF, Research Report 2007-03)
Executable and source code for Windows 2000/XP/Vista
(believed to be usable under Mac OS X and Linux by same methods as CPIDR 3; see manual)

CPIDR® (Computerized Propositional Idea Density Rater) version 3.2 for Windows
Documentation (PDF, Research Report 2007-03)
See also the published paper listed below.
CPIDR 3.2 installation package for Windows 2000/XP/Vista/7/8 (zipped MSI)
Alternate installation package for CPIDR 3.2 for Windows (zipped MSI; built with newer version of compiler, but functionally identical)
Running CPIDR under Mac OS X, UNIX, and Linux (PDF, Research Report 2007-04)
NOTE: We have had reports that the Mac clipboard does not sync with the CPIDR clipboard, and that CPIDR crashes on the Macintosh under conditions that we have not pinned down. Please remember that this is Windows software, and running it under other operating systems is experimental.
CPIDR 3.2 files for Mac OS X, UNIX, Linux (zipped package of individually accessible files)

CPIDR 5, an improved commercial (non-GPL) version of CPIDR, is free for noncommercial users, though not open-source. See the documentation to learn how to obtain a commercial license.

CPIDR 5 can be called (as a DLL) from other software, and we encourage licensing it for incorporation into commercial products.

Download documentation (PDF)
Download software (for Windows Vista, 7, 8)

The name CPIDR is a registered trademark of The University of Georgia Research Foundation, Inc.

Note: Both versions of CPIDR can analyze multiple text files in a single run. Simply select multiple files when prompted to "Choose File(s)."

An idea density measurement program for French, Densidées, modeled on CPIDR, is being developed by Hyeran Lee and others and is available, free, from http://code.google.com/p/densidees/.

ShaC (Shallow Syntactic Complexity Analyzer)
Prolog source code, data files, and user's guide
Research report 2005-01


Published Papers (most recent first)

Elvevåg (Elvevaag), Brita; Wynn, Rolf; and Covington, Michael A. (2011) Case report: Meaningful confusions and confusing meanings in communication in schizophrenia. Psychiatry Research 186:461-464. PDF.

Covington, Michael A. and McFall, Joe D. (2010) Cutting the Gordian knot: the moving-average type–token ratio (MATTR). Journal of Quantitative Linguistics 17:94-100. Full text on InformaWorld

Covington, Michael A. (2009) Idea density: a potentially useful characteristic of retrieved documents. Proceedings, IEEE SoutheastCon 2009. Presentation slides (PDF)

Nicholson, Colin (2009) Judging whether a document changes in subject. Proceedings, IEEE SoutheastCon 2009.

Covington, Michael A.; Riedel, Wim J.; Brown, Cati; He, Congzhou; Morris, Eric; Weinstein, Sara; Semple, James; Brown, John (2009) Ketamine and schizophrenic speech: more difference than originally reported. Journal of Psychopharmacology 23 (1) 111-112.

Brown, Cati; Snodgrass, Tony; Kemper, Susan J.; Herman, Ruth; and Covington, Michael A. (2008) Automatic measurement of propositional idea density from part-of-speech tagging. Behavior Research Methods 40 (2) 540-545.

Covington, Michael A.; Riedel, Wim J.; Brown, Cati; He, Congzhou; Morris, Eric; Weinstein, Sara; Semple, James; Brown, John (2007) Does ketamine mimic aspects of schizophrenic speech? Journal of Psychopharmacology 21 (3) 338-346.

Covington, Michael A.; He, Congzhou; Brown, Cati; Naci, Lorina; McClain, Jonathan T.; Fjordbak, Bess Sirmon; Semple, James; Brown, John (2005) Schizophrenia and the structure of language: the linguist's view. Schizophrenia Research 77(1):85-98, 2005.


Published Abstracts (most recent first)

Vin, Hemali (2009) Measuring the orderliness of thought. University of Georgia Center for Undergraduate Research Opportunities (CURO) Symposium, April 6, 2009.
Abstract (PDF)

He, Congzhou; Weinstein, Sara; and Covington, Michael A. (2007) Using text analysis software in schizophrenia research. Schizophrenia Bulletin 33:522.

Brown, Cati; Covington, Michael A.; Semple, James; and Brown, John (2005) Reduced idea density in speech as an indicator of schizophrenia and ketamine intoxication. Schizophrenia Bulletin 31:187-188.

He, Congzhou; Covington, Michael A.; Semple, James; and Brown, John (2005) Some linguistic signs of ketamine-induced cognitive impairment. Schizophrenia Bulletin 31:511.


Presentations (most recent first)

Covington, Michael A.; Lunden, Anya; Cristofaro, Sarah; Johnson, Stephanie; Ramsay, Claire; Broussard, Beth; Zhang, Shayi; Bailey, C. Thomas; Fogarty, Robert; and Compton, Michael T. Phonetic measurement of reduced facial muscle movement among young adults with first-episode schizophrenia-spectrum disorders, poster presented at Georgetown University Round Table on Languages and Linguistics, Washington, D.C., March 2012.

He, Congzhou; Weinstein, Sara; and Covington, Michael A. Using text analysis software in schizophrenia research, poster presented at International Congress on Schizophrenia Research, Colorado Springs, March 2007.

Brown, Cati; Snodgrass, Tony; Covington, Michael A.; Herman, Ruth; and Kemper, Susan J., Measuring propositional idea density through part-of-speech tagging, poster presented at Linguistic Society of America Annual Meeting, Anaheim, California, January 2007.

Baptista, Marlyse; Fjordbak, Bess; and Covington, Michael A., Linguists meet clinicians: a study of language disorders, poster presented at the University of Georgia Linguistics Conference, Athens, Georgia, September 2006.

He, Congzhou; Weinstein, Sara; and Covington, Michael A., Speech analysis software for psychiatric research: the case of D-Level Rater, poster presented at the First annual Georgia/South Carolina Neuroscience Consortium, Charleston, April 2006.

Brown, Cati; Covington, Michael A.; Semple, James; and Brown, John, Reduced idea density in speech as an indicator of schizophrenia and ketamine intoxication, poster presented at the International Congress on Schizophrenia Research, Savannah, April 2005.

He, Congzhou; Covington, Michael A.; Semple, James; and Brown, John, Some linguistic signs of ketamine-induced cognitive impairment, poster presented at the International Congress on Schizophrenia Research, Savannah, April 2005.

He, Congzhou; Brown, Cati; Covington, Michael A.; and Naci, Lorina, How complex is that sentence? A proposed revision of the Rosenberg and Abbeduto D-Level scale, poster presented at the annual meeting of the Linguistic Society of America, Boston, January 2004.


CASPR Research Reports (most recent first)

Research Report 2007-05
Michael A. Covington
MATTR User Manual

Research Report 2007-04
Cody Boisclair
Running CPIDR under Mac OS X, UNIX, and Linux

Research Report 2007-03
Michael A. Covington
CPIDR 3 User Manual

Research Report 2007-02
Michael A. Covington and Joe D. McFall
Using MontyLingua 2.1 with C# and Microsoft .NET

Research Report 2007-01
Colin Nicholson
Schizophrenia and the Structure of Language: Annotated Bibliography 2004-2006

Research Report 2006-02
Congzhou He
Using Text Analysis Software in Schizophrenia Research

Research Report 2006-01
Michael A. Covington, Congzhou He, Cati Brown, Lorina Naci, and John Brown
How Complex is That Sentence? A Proposed Revision of the Rosenberg and Abbeduto D-Level Scale

Research Report 2005-02
Salena A. Sampson
Computerized Analysis of Salient Items and Discourse Comprehension in Schizophrenic Picture Descriptions

Research Report 2005-01
Matthew J. Voss
Determining Syntactic Complexity Using Very Shallow Parsing
Computer program (as Prolog source code), files, and documentation (ZIP)

For research reports of the Institute for Artificial Intelligence (not CASPR), click here.


The former CASPR internal web site is no longer accessible by Web; use VPN or SFTP.



How to make a PDF file of a printed document using Acrobat 8 — read this if you are involved in maintaining the CASPR library.


Obligatory training for new researchers:

New members of the CASPR team must complete an online training course in human subjects research ethics. This does not have to be done all at once; it is divided into 30-minute modules. It is required if you are going to work with human subjects data or if you are going to be paid from an NSF grant.

To receive the training, go to http://www.ovpr.uga.edu/hso/training/ and follow the instructions. Choose either the Social and Behavioral or the Bio-Medical course. When you finish, print out and save a copy of your certificate, and also e-mail the principal investigator.