This thesis examines information-processing in chronic pain. "Schematic" processing is investigated selective memory for pain-related information is explored in depressed and non-depressed chronic pain patients, depressed non pain-patients and controls. A memory bias for sensory adjectives is found in the non depressed chronic pain group, while a tendency to over-recall both sensory and affective compared to neutral information is found in the depressed chronic pain group. No memory bias is observed in an acute pain group, and the implications of this are discussed. A possible cognitive avoidance mechanism is identified in depression. A questionnaire assessing beliefs about pain ("conceptual" processing) is developed and validated, and shown to differentiate between chronic pain patients and controls. The impact of two interventions for chronic pain (surgery and cognitive-behavioural management) on schematic and conceptual processing is investigated prospectively. In general the endorsement of organic beliefs decreases while the emphasis on psychological beliefs increases post-intervention. Evidence is found to suggest that surgery, but not cognitive-behavioural treatment, reverses pain-related memory biases. This is discussed in relation to changes in pain intensity. Evidence is provided to suggest that beliefs are causally related to several pain-related measures including anxiety, depression, health locus of control, cognitive coping strategies and activity levels. A word completion paradigm is employed to explore further the role of schematic processing in chronic pain, and finally, a lexical decision task is used to assess the role of word frequency effects in information-processing in chronic pain. These results suggest that memory biases in chronic pain cannot be explained by frequency effects, hence addressing the validity of the memory biases described earlier in the thesis.