Uncomfortable bedfellows; according to Bishop this constitutes the relationship between cognitive neuropsychology and developmental disorders. Boyle et al. (2011) showed that one out of six children suffer from a developmental disorder. The ICD-10 (WHO, 1992) defines developmental disorders (DD) as fulfilling three categories: an early onset, a delay in developmental functions relating to the CNS or biological maturation and a lack of remission. The DSM-IV (APA, 2000) takes a different direction by categorising Pervasive Developmental Disorders as characterized by impaired development of reciprocal social interaction and communication skills and the presence of stereotyped behaviour, interests, and activities. Whilst the first definition describes DD like developmental Prosopagnosia (face blindness, DP) which affects around 2.47% of the population (Kennerknecht, Grueter T., Welling, Wentzek, Horst, Edwards, & Grueter M., 2006), the latter definition matches the triad of impairments of autism (ASD) which has a prevalence of 1% (Boyle et al., 2011). Comorbidity of DD such as ASD and prosopagnosia can result (Bate, 2012). At the same time, around 20% of individuals over 18 years suffer from acquired brain disorders and diseases (Family Caregiver Alliance, 2012). Deficits in acquired and developmental individuals are commonly mapped out using functional models and explored via double-dissociations. These are usually derived from acquired disorder (AD) patients rather than DD patients. How these factors influence the debate of cognitive neuropsychology and DD making “uncomfortable bedfellows” shall be explored in the following.
Several points will be discussed as having an influential role in the debate of cognitive neuropsychology (CN) accounting for developmental disorders (DD). These are: modularity, plasticity and compensation, competence and performance, differences between exploring AD and DD and the role of double-dissociations in relation to developmental versus acquired disorders as displayed in CN.
The association between structural and cognitive development is gaining recognition. Developmental changes in task-specific processing such as that of faces have been argued to play an important role in the amalgamation of various brain networks and experience-based growth of neural architecture (Westermann, Sirois, Shultz, & Mareschal, 2006). Also, a bidirectional interaction between behavioural and neuronal development has been established as learning supposedly influences brain structuring and vice versa. Similarly, the rate of neurogenesis arguably depends on learning and environmental circumstances (Westermann et al., 2006).
(Bishop, 1997, p.902)
Following on from that, cognition is executed by the growing neurofunctional network. This distributed parallel processing arguably is organised into interacting modules (Table 1), which comprise of different levels of complexity depending on the cognitive task or function carried out (Jackson, 1958). This can be illustrated by “Theory of Mind” (Baron-Cohen, 1995) in ASD: ToM proposes developmental impairment of the ability to appreciate own and others’ mental states (beliefs, desires or intentions etc).
Adams (2010), Brothers’ (1990) theory of the Social Brain and Leslie’s (1991) ToM Mechanism argue that key features of ToM can be explained in terms of modularity. Marchery (forthcoming) supports this as modularity poses a crucial characteristic of biological development. Arguably, deficits are not due to minor developmental brain problems in lower-level input and consequential cascading abnormal interaction between input and higher-level systems that affect all cognitive capacities (Karmiloff-Smith, 1998; 1992; Machery, forthcoming).
Nonetheless, this is what Gerrans P.,T. and Stone (2008) argue. Additionally, low-level deficits such as gaze and face processing prevent ASD individuals from having the same experience as typically developing individuals, caused by a deficit in domain-general capacity. This has the aforementioned cascading interaction effect during development rather than leading to a dysfunctional mindreading module. However, Gerrans et al. do not see development as single causality.
Reconciling, Baron-Cohen (1995) postulates that ToM is either seen as innate module or explained by deficits in general learning mechanisms and unifies both in the intermediate model of Social Perception of Minimalist Innate Modularity.
(Amaral, Schumann, & Nordahl, 2008, p.2)
Figure 1 displays brain structures associated with dysfunctions in ASD. Amaral et al. (2008) join the debate of modularity suggesting that despite functional identification of modules no clear pathology has emerged yet but rather, that e.g. Schumann and Amaral, (2004) have demonstrated that the developmental trajectory itself rather than the final product exhibits most disruption and by this posing a further point against applying CN to the explanation of DD.
Functions of modularity imply, on the other hand, that one zone can belong to several overlapping neuronal networks (Damasceno, 2010).Disordered brain development can lead to structural disruption and thereby to several dysfunctions which may not appear as a specific deficit but as a syndrome (a collection of symptoms) as demonstrated in the triad of impairments in ASD (Wing & Gould, 1979).
Similarly, lesions to a localised brain area do not always cause isolated but sometimes collective symptoms (Kolb, Brown, Witt-Lajeunesse, & Gibb, 2001). These are caused, i.a. by disturbances in interconnected regions which might appear intact on functional brain imaging scans. However, changes in neurotransmission, excitatory or inhibitory processes or blood flow might cause secondary symptoms not predicted by the data. One interesting example was proposed by Sandel, Weiss and Ivker (1990): their patient was described as amnesic after traumatic brain injury. Her impaired cognition was considered a secondary deficit as her EEG and other clinical tests proved negative. No other disruptions were present until she was eventually diagnosed with multiple personality disorder.
Therefore several unexpected and seemingly unconnected functional disruptions might result from a single localised lesion.
These points argue against the notion of using CN in explaining DD due to the assumptions held.
Advancing to the next argument, functional (mal)development following lesions or disrupted neurogenesis might also demonstrate individual differences in brain structure, due to varying input, demands and circumstances. Thomas and Karmiloff-Smith (2003) added that specialisation/modularity is not directing development but contrarily, is the outcome of this. On the other hand, the brain activity of healthy individuals supports a “normalisation process” to facilitate models being created for generalisation purposes.
As Tager-Flusberg (1999) counter-argues, functional variation in dysfunctional individuals is similar to that of typical population. Allegedly, localisation is universally alike thereby plasticity, compensation or adaptation to damage are neglected. Examples of compensatory processes and resulting behaviour (or v.v.?) are stimming in ASD for sensory overload (Grandin, 2009) or remembering haircuts in prosopagnosics for identification (Bate, 2012).
Adaptation can express itself by change in neurochemistry, psychomotor stimulants and compensatory behaviour while it is argued that the behaviour triggers alteration in neurochemistry (Kolb et al., 2001). This in turn, points to the aforementioned bidirectionality of structure and learning which has implications for acquired as well as developmental patients and goes against explaining DD in a CN framework.
At this point, double-dissociations are commonly applied to provide clues to brain-behaviour relationships. In 1955, Teuber introduced the term double-dissociation referring to mirror-deficit patients. Buchaine, Parker, and Nakayama (2003) proposed the example of NM, a prosopagnosic whose abilities support previous findings (Young, Newcombe, De Haan, Small, & Hay, 1993) of a double-dissociation of identity and emotion recognition. Van Peelen, Lucas, Mayer, and Vuilleumier (2009) added that emotional guidance of attention to facial expression can be unimpaired and thus functionally differ from face identification.
Interestingly, Teuber did not establish double-dissociations to discover networks within cognitive architecture (Van Orden, Pennington, & Stone, 2001) as localising a lesion does not imply a responsibility for dysfunctions.
Further, modularity of functional networks has not been unanimously established. Nonetheless, modularity is one key feature of CN serving the notion of double-dissociations. CN informs theories of normal cognitive function by analysing the break down of cognitive systems in atypical patients (Bullinaria & Chater, 1997). However, inferring theory from damage patients is full of obstacles and depends on models of normal functioning. As Caramazza (1984) argued, double-dissociations are valid forms of inference but only if the cognitive model is well developed; only single-case studies can provide information relevant to our understanding of cognitive architecture.
Contradicting, Shallice (1988) argued that a double-dissociation between two tasks does not necessarily imply a double-dissociation between cognitive processes. Since the argument concerning emergence of CN models and the role of AD and DD has not been resolved yet, a need of inclusion of further aspects becomes apparent.
One aspect is Residual Normality which claims that atypical development impairs only specific domains while the rest functions typically (Thomas & Karmiloff-Smith, 2002). Similar behavioural impairments demonstrated by adult brain damage and infant-onset disorders are assumed to support modularity of typical neuronal system structure.
Opposing the assumption of RN, compensation due to damage causes alterations in various other modules (Thomas, 2003). Shallice (1988) introduced deriving knowledge about the neuronal system from behavioural impairments in AD and DD, such as malfunctioning ToM in autism (Leslie, 1991). As argued by Bishop and Karmiloff-Smith (1997; 1998) such an inferential process within the static model of CN leads to invalid explanations of dysfunctional development.
Functional Models on the other hand do not possess a developmental component, and display what is expected to happen under various conditions in terms of functionality and processes (D’Mello & Franklin, 2011). Thus impairments can be analysed under the assumption of RN as it has been done, for example in the study of autism and prosopagnosia (Thomas et al., 2002). Examples of models are given in Figure two and three. Despite this, it has been suggested to use connectionist models where in AD the damage is applied at the end of training and for DD prior to it as this is argued to reflect actuality in a truer light. Connectionist models indeed are suitable for the investigation of DD as behavioural deficits such as in autism (Thomas et al., 2002) can be captured well following simulated atypical development. It is highlighted that AD and DD are studied using separate models, thereby arguing against functional models within CN.
Figure 2: Functional model of face-processing (Wright, Wardlaw, Young, & Zeman, 2006)
Figure 3: Models of Autism (Anderson and Herbert, 2008)
The aforementioned factors influencing whether cognitive neuropsychology is suitable for explaining DD have given a mixed picture. Jackson and Coltheart (2001) defended the use of CN in DD stating that development itself (thus distal causes like genes or background) are irrelevant for the identification of un/healthy functions, given modularity. Whether acquired or developmental – proximal (what is wrong right now) and distal causes can be deduced independently from each other. Indeed, Jackson et al. argue that common architectural issues in AD and DD can be established using CN.
Bishop on the other hand postulates against this as deficits are focused on while intact processes are neglected. Further, top-down and bottom-up interactions are dismissed. Karmiloff-Smith (1998) deems these highly relevant for understanding DD. Thirdly a uniform modularity is assumed without accounting for alterations in development. Bishop proposed that AD should be researched using single-case double-dissociations while for DD, as they are likely to demonstrate associated deficits, group studies are recommended.
Problematically, the debate on what constitutes a module has not been clarified yet (e.g. Fodor (2000): encapsulation, Coltheart (1999): domain specificity). Until this is established functional models and CN will not carry valid explanations nor reveal (mal)development well (Thomas et al., 2002). This underpins the view against exploring DD within a CN framework.
Additionally, modularity/plasticity’s scope of variation in DD and AD remains unestablished (Tager-Flusberg, 2000; Thomas et al., 2003). Humphreys, Ewing, and Karmiloff-Smith (2002) proposed the example of Williams syndrome patients possessing intact face recognition which is, however, achieved by applying processes diverging from typicality. Initially, face recognition was seen as intact despite visuospatial deficits, reduced sensitivity to inverted faces and lack of progressive development of localisation (Grice, Spratling, Karmiloff-Smith, Halit, Csibra, De Haan, & Johnson, 2001). Hypothetically, assuming RN, this could be seen as double-dissociation between developmental prosopagnosia and WS. Basing modelling of DD on highly debated assumptions poses a problem for validity of the framework.
Additionally, CN is accused of focusing on representational (competence) rather than processing (performance) deficits (Bishop, 1997). This, however, appears to be crucial in terms of applying research to DD. Weigelt, Koldewyn, and Kanwisher (2012) demonstrated a lack of qualitative difference (how facial identity is discriminated or remembered, competence) but rather a quantitative (how well, performance) difference in that ASD patients are impaired in memory and perception despite indication of intact face identity recognition. According to Bishop, this is another argument against applying CN to exploring DD.
Concluding then, the much debated assumption of modularity held in cognitive neuropsychology overshadows exploration of disruptions during the trajectory of neuronal development. Applying double-dissociations to DD arguably yields invalid results and is more suited to acquired disorders. This is due to the assumptions within CN of Residual Normality, directionality of impact within disorders and its dependence on localisation of disruption, universality of localisation, neglect of plasticity and the focus on (in)competence. All in all, it has been proposed, while certain aspects could be overlooked, connectionist models rather than functional models would yield better results. Also, cognitive neuropsychology focuses on adult processes thus damage after complete development which is hard to match onto disorders of development. In order for CN to account validly for disorders, different models have to be applied for acquired and developmental cases.