CDP : The Connectionist Dual Process model of reading aloud
What is CDP ?
The Connectionist Dual Process (CDP) model is the leading computational model of reading aloud for English, Italian, French and German. It is based on a connectionist dual route architecture first described by Zorzi, Houghton & Butterworth (1998, JEP:HPP), in which the emergent "division of labor" between lexical and nonlexical processing is grounded in the different computational properties of the neural networks that implement these processes. The direct spelling-to-sound pathway responsible for phonological assembly (TLA network) is the key to the model's success in accounting for a wide body of empirical data from studies on skilled oral reading, learing to read, and reading disorders (dyslexia) (for a review see Zorzi, 2010, EJCP). The connectionist dual route architecture has also been successfully used to develop a computational model of spelling (Houghton & Zorzi, 2003, Cognitive Neuropsychology)
Three major upgrades have followed the original version of the English CDP model (Zorzi, Houghton & Butterworth, 1998):
- CDP+ (Perry, Ziegler, & Zorzi, 2006, Psychological Review), with a fully implemented lexical route and a grapheme level of orthographic representation (graphemic buffer) in the sublexical route. CDP+ is the result of an incremental and nested modelling strategy, according to which a new model should build upon the strengths of previous models while eliminating their weaknesses. At present, CDP+ is the most successful computational model of reading monosyllabic English words aloud and it beats all previous models by an order of magnitude when predicting individual item-level variance on large databases.
- CDP++ (Perry, Ziegler, & Zorzi, 2010, Cognitive Psychology), which extends the model to English bisyllabic word and nonword naming, including stress assignment. The model has a lexicon of more than 32,000 words and it can explain up to 49% of the reaction time variance at the item level in large scale databases like the English Lexicon Project.
- CDP++.parser (Perry, Ziegler, & Zorzi, 2013, Cognitive Science), extends CDP++ with a learned graphemic parsing mechanism. The latter is modeled with a simple network that learns to parse the letter string into graphemes that are also categorized as onset, vowel or coda.
CDP models in other languages
- German CDP+ (Perry, Ziegler, Braun, & Zorzi, 2010, European Journal of Cognitive Psychology).
- French CDP++ (Perry, Ziegler, & Zorzi, 2014, Journal of Memory and Language). Multisyllabic model (up to three syllables) with a lexicon of more than 100,000 words, it explains up to 66% of item-specific RT variance in large databases
- Italian CDP++ (Perry, Ziegler, & Zorzi, 2014, PLoS ONE). Multisyllabic model (up to three syllables and stress assignment) with a lexicon of more than 60,000 words.
Reading development and dyslexia in CDP
In Ziegler, Perry & Zorzi (2014, Phil. Trans. Royal Soc. B) we implement and test the developmentally plausible phonological decoding self-teaching hypothesis in the context of the connectionist dual process model. In a series of simulations, we show that the model can acquire word-specific orthographic representations for more than 25 000 words even though it starts with only a small number of grapheme–phoneme correspondences. We then show how visual and phoneme deficits that are present at the outset of reading development can cause dyslexia in the course of reading development.
Download and use CDP
CDP, in its various versions, is available to other researchers as stand-alone MS-Windows program. Each zip file includes the program files for the specific version and a user manual. No installation is required (just extract all files to the same folder). Researcher who would like to use the same item sets (for CDP+, including benchmark effects) for their own simulation studies can download all stimuli (and behavioral data) from here.
- CDP+ can be downloaded here.
- CDP++ can be downloaded here.
- French CDP++ can be downloaded here.
- Italian CDP++ can be downloaded here.
Questions or comments about CDP
References for CDP+
Main references to the CDP work:
- Perry, C., Ziegler, J.C., & Zorzi, M. (2006). Nested incremental modeling in the development of computational theories: The CDP+ model of reading aloud. Psychological Review, 114, 273-315
- Perry, C., Ziegler, J. C., & Zorzi, M. (2010). Beyond single syllables: Large-scale modeling of reading aloud with the Connectionist Dual Process (CDP++) model. Cognitive Psychology,61, 106-151
- Perry, C., Ziegler, J. C., & Zorzi, M. (2013). A computational and empirical investigation of graphemes in reading. Cognitive Science, 37, 800-828.
- Perry, C., Ziegler, J.C., & Zorzi, M. (2014). When silent letters say more than a thousand words: An implementation and evaluation of CDP++ in French. Journal of Memory and Language, 72, 98–115..
- Perry, C., Ziegler, J.C., & Zorzi, M. (2014).CDP++.Italian: Modelling sublexical and supralexical inconsistency in a shallow orthography. PLoS ONE 9(4): e94291.
- Perry, C., Ziegler, J.C., Braun, M., & Zorzi, M. (2010). Rules versus statistics in reading aloud: New evidence on an old debate. European Journal of Cognitive Psychology, 22, 798-812.
- Ziegler, J. C., Perry, C., & Zorzi, M. (2014). Modelling reading development through phonological decoding and self-teaching: Implications for dyslexia. Philosphical Transactions of the Royal Society B, 369: 20120397.
- Zorzi, M. (2010). The Connectionist Dual Process (CDP) approach to modeling reading aloud. European Journal of Cognitive Psychology, 22, 836-860.
- Zorzi, M., Houghton, G., & Butterworth, B. (1998). Two routes or one in reading aloud? A connectionist dual-process model. Journal of Experimental Psychology: Human Perception and Performance, 24, 1131-1161.
Other relevant publications
Below is a list of our other CDP-related articles.
- Ziegler, J.C., Perry, C., & Zorzi, M. (2009). Additive and Interactive Effects of Stimulus Degradation: No Challenge for CDP. Journal of Experimental Psychology: Learning, Memory and Cognition, 35, 306-311.
- Zorzi, M. (2005). Computational models of reading. In G. Houghton (Ed.), Connectionist models in Cognitive Psychology (pp. 403-444). London: Psycholoy Press.
- Hutzler, F., Ziegler, J., Perry, C., Wimmer, H., & Zorzi, M. (2004). Do current connectionist models account for reading development in different language? Cognition, 91, 273-296.
- Houghton, G., & Zorzi, M. (2003). Normal and impaired spelling in a connectionist dual-route architecture. Cognitive Neuropsychology, 20,115-162 .
- Perry, C., & Ziegler, J. C. (2002). Cross-language computational investigation of the length effect in reading aloud. Journal of Experimental Psychology: Human Perception and Performance, 28, 990-1001.
- Zorzi, M. (2000). Serial processing in reading aloud: No challenge for a parallel model. Journal of Experimental Psychology: Human Perception and Performance, 26, 847-856.
- Zorzi, M., Vigliocco, G. (1999). Dissociation between regular and irregular in connectionist architectures: Two processes, but still no special linguistic rules. Behavioral and Brain Sciences, 23, 1045-1046.
- Zorzi, M. (1999). Routes, races, and attentional demands in reading: Insights from computational models. In: M. Hahn & S.C. Stoness (Eds.), Proceedings of the Twenty First Annual Conference of the Cognitive Science Society (p. 772-777). Mahwah (NJ): Erlbaum.
- Zorzi, M., Houghton, G., & Butterworth, B. (1998). The development of spelling-sound relationships in a model of phonological reading. Language and Cognitive Processes, 13, 337-371. (Reprinted in K. Plunkett (Ed.), Language Acquisition and Connectionism. Hove: Psychology Press.)
- Houghton, G., & Zorzi, M. (1998). A model of the sound-spelling mapping in English and its role in word and nonword spelling. In M.A. Gernsbacher (Ed.), Proceedings of the Twentieth Annual Conference of the Cognitive Science Society (p. 490-501). Mahwah (NJ): Erlbaum.