Use this URL to cite or link to this record in EThOS:
Title: Parallel Distributed Processing (PDP) models as a framework for designing cognitive rehabilitation therapy
Author: Nte, Solomon
ISNI:       0000 0004 5353 265X
Awarding Body: University of Manchester
Current Institution: University of Manchester
Date of Award: 2015
Availability of Full Text:
Access from EThOS:
Access from Institution:
Parallel Distributed Processing (PDP) modelling has simulated developmental learning across a range of domains such as reading (e.g. Seidenberg & McClelland,1989) or Semantics (e.g. Rogers et al. 2004). However aside from two notable exceptions (Plaut, 1996; Welbourne & Lambon Ralph, 2005b) modelling research has not addressed the simulation of relearning during spontaneous recovery or rehabilitation after brain damage, and no research has considered the effect of the learning environment. This thesis used an established PDP model of semantic memory (Rogers et al., 2004) to simulate the influence of the learning environment. A novel quantitative measure (called representational economy) was developed to monitor efficiency during learning. Developmental learning is considered to be multimodal (e.g. Gogate et al., 2000) whereas rehabilitation is normally carried out through therapy sessions employing unimodal learning tasks (Best & Nickels, 2000). This thesis hoped to discover whether multimodal rehabilitation may be more efficient (as suggested by Howard et al., 1985). Three sets of simulations were conducted: The first set contrasted multimodal and unimodal learning in development and recovery, and tested internal representations for robustness to damage finding multimodal learning to be more efficient in all cases. The second set looked at whether this multimodal advantage could be approximated by reordering unimodal tasks at the item level. Findings indicated that the multimodal advantage is dependent upon simultaneous item presentation across multiple modalities. The third set of simulations contrasted multimodal and unimodal environments during rehabilitation while manipulating background spontaneous recovery, therapy set size and damage severity finding a multimodal advantage for all conditions of rehabilitation. The thesis findings suggest PDP models may be well-suited to predicting the effects of rehabilitation, and that clinical exploration of multimodal learning environments may yield substantial benefits in patient-related work.
Supervisor: Not available Sponsor: Not available
Qualification Name: Thesis (Ph.D.) Qualification Level: Doctoral
EThOS ID:  DOI: Not available
Keywords: Computational Model ; Speech and Language Therapy