Future Directions: Can Neuroscience Contribute to the Study of Cognitive Modification?

Nicholas Hon

In the last few years, it has become increasingly apparent that the neural bases of cognitive systems can be fruitfully studied. A natural question arising from this development is whether studying the brain can offer insight into how cognitive systems may be modified. This chapter considers various ways in which neuroscience can inform the study of cognitive modification.

Change or modification is a notion that holds fascination for many different fields of inquiry. In psychology, we are fascinated by the idea that our cognitive processes may be modified, for enhancement or perhaps remediation. In neuroscience, a similar fascination can be found; however, it is modification of the brain (in terms of either structure or function) that is studied. Although not a fully developed field of research, given the tight coupling between brain and mind, there has been a steadily growing interest in the possibility that neuroscience might be able to augment or extend traditional methods of inquiry into cognitive modification (e.g., Ansari & Coch, 2006; Goswami, 2004; Posner & Rothbart, 2005). In this chapter, by way of example, three ways in which neuroscience, particularly brain imaging, can contribute to the study of cognitive modification will be discussed.

IDENTIFYING THE “CORRECT” NEURAL NETWORKS

The advent and improvements to brain imaging technology has allowed, perhaps for the first time, researchers to consider the neural substrate of different cognitive operations. Methods like functional magnetic resonance imaging (fMRI) and magnetoencephalography (MEG) allow for neural activity associated with mental processing to be assessed safely in humans. Findings from these methods can be combined with data from invasive cell-recordings from primates and other animals to create a picture of how the brain supports the mind at many different levels, from the anatomical to the neuronal. One obvious use of brain imaging techniques would be in the identification and understanding of the neural circuits that underlie the cognitive systems that we are trying to modify. Cognitive systems are supported by neural systems. A proper understanding of these neural systems would allow for greater insight into those aspects of a cognitive system that can be readily modified and those not so easily modified. It can also give us a benchmark by which we can monitor the success of an intervention.

Consider the case of reading, which is a complex mental activity that requires the coordination of different mental operations. When reading a printed word, individual letters must be chunked into a word, phonological information must be decoded and meaning must be extracted. Correspondingly, brain imaging studies have determined that skilled reading requires the engagement of many different parts of the brain including occipital, temporal, and prefrontal cortex. Each of these areas appears to play a different role in reading. For example, an occipitotemporal area appears to be related to the processing of visual word forms (Cohen et al., 2002), whereas an inferior prefrontal area appears to be involved in accessing semantic information (Wagner et al., 2001). Skilled reading requires that all of these operations and their supporting neural substrates be intact and uncompromised. Studies have also shown how activity of these areas is related to reading behavior. For example, when subjects have to access more, relative to less, semantic information, greater activation of left inferior frontal cortex is observed (Wagner et al., 2001). Also, reading ability appears to be related to the appropriate activation of this network. Poorer readers demonstrate less activation of this “reading network” than more skilled readers (Shaywitz et al., 2004). Correspondingly, dyslexia, a reading disorder, is associated with underactivation of the reading network (Temple et al., 2003).

Given knowledge of the “correct” neural circuit serving a particular cognitive operation, an appropriate enhancement package should aim to enhance by altering the working of this same network. Again, let us consider the case of reading. Research has suggested that facility with phonemes (or phonological awareness) is important in learning to read (Stanovich, Cunningham, & Cramer, 1984). A recent study attempted to enhance reading performance by providing children with a phono-logically-based instruction program (Shaywitz et al., 2004). Reading fluency was improved for children who were given this instruction, and this improvement extended to comprehension. fMRI scans revealed these children showed increased frontal and occipito-temporal activation post-training. As discussed earlier, these brain areas are known to support reading. This finding suggests that the training worked by influencing the “normal” reading network and not by recruiting other peripheral neural resources. Research has shown that, for the same task, performance can be similar but be based on engagement of different neural resources. For example, when bilinguals equally proficient with both languages performed a phonological task, they leveraged on different neural resources than “unequal” bilinguals (Chee et al., 2004). Possibly, this difference in neural recruitment pattern underlies some of the differences in linguistic facility between the two groups, suggesting that the correct neural resources need to be engaged for optimal or better performance.

Knowledge of the “normal” networks underlying a cognitive operation can also provide hints on how to assess remediation programs. Remediation is an attempt at modification because the current state of a cognitive system is somehow deficient. As suggested previously, phonological awareness appears to be important in skilled reading. Training focused on phonological awareness given to dyslexic children resulted in greater activation of areas of the brain related to phonological processing in skilled readers (Temple et al., 2003), which suggests that the remediation program was targeting the “correct” areas. This increase in activation was observed in tandem with an increase in performance on tests of naming skills and comprehension.

BRAIN IMAGING OFFERS INSIGHT INTO HOW COGNITIVE SYSTEMS WORK

The previous section highlighted the usefulness of brain imaging in identifying the canonical neural circuits that underlie cognitive systems. However, on occasion, the judicious combination of brain imaging and behavioral measures can even provide insight into the psychological mechanisms of a cognitive system. One good of example of this relates to numerical ability. A recent combined behavioral and brain imaging study demonstrated that numerical thought may comprise different components (Dehaene et al., 1999). English-Russian bilinguals were taught to perform exact addition (e.g., 4 + 5 = 9) and approximation tasks (e.g., 4 + 5 is closer to 8 than to 2) in one of the two languages they were familiar with. When subjects were tested on these tasks, it was observed that performance in the exact addition task was better when the task was presented in the teaching language. On the other hand, performance in the approximation task was the same regardless of which language the test was conducted in. This same pattern of results was observed even when the tasks involved more complex mathematical operations (e.g., involving cube roots and logarithms). The fMRI data collected while subjects performed the tasks indicated that exact arithmetic engaged a largely left-lateralized network including left inferior frontal cortex, an area known to be involved in linguistic processing. This supports the idea that exact arithmetic may leverage on language-dependent representations. On the other hand, approximation engaged bilateral parietal regions. Parietal areas are known to be involved in spatial processing. Therefore, one possibility is that approximation may involve representations of numerical magnitude that are analogous to representations of spatial magnitude or distance.

What makes this finding interesting is that it suggests that an important part of numerical thinking may be linked to the working of the linguistic system. What are the implications of this? Often, training for mathematics is conducted independent of linguistic training, perhaps on the assumption that they are fundamentally different cognitive operations. Educators may find it worthwhile to consider the possibility that systematic programs augmenting traditional methods of mathematical training with a linguistic component may enhance numerical facility. It also suggests the importance of emphasizing visuo-spatial information in the training of numerical skills. One obvious example of this would be the graphical depiction or representation of numerical data. Therefore, practitioners who are attempting to devise modification programs targeting the “numeracy system” may find it useful to consider both “streams” and their interaction when designing their programs.

NEUROSCIENCE CAN OFFER SOME NOVEL IDEAS

This section discusses an intriguing neuroscience finding which might offer a new line of inquiry for the study of cognitive modification.

Recently, it was observed that diverse mental demands appear to leverage on common neural resources (Cabeza & Nyberg, 2000; Duncan, 2006; Duncan & Owen, 2000). For example, when the peak activations obtained from five different classes of cognitive demands (response conflict, task novelty, working-memory load, working-memory delay, and perceptual difficulty) were compared, it was noticed that the different cognitive tasks activated similar (if not identical) parts of the brain, specifically in frontal and parietal cortex. Common frontal activations included frontal operculum, inferior frontal sulcus and anterior cingulate cortex. It was also observed that these demands all seemed to rely on the engagement of the posterior parietal lobe. Notice that the different types of demands or tasks considered in that meta-analysis were very different; for example, perceptual difficulty tasks involved making decisions or discriminations under conditions of perceptual degradation whereas working-memory-load tasks involved maintaining information in working or short-term memory. Nor is this pattern of activity only restricted to only these five types of cognitive demand. It has been observed in other cognitive domains. The same specific frontal and parietal activation is observed, for example, in studies manipulating aspects of language (Jiang & Kanwisher, 2003) and semantic memory (Wagner et al., 2001), as well as when people are engaged in planning behavior (Fincham et al., 2002).

This finding is particularly interesting because it suggests that diverse cognitive demands may leverage on a common body of neural resources. The finding that similar areas are activated by different tasks goes some way in suggesting that these areas may support a general function common to many different tasks and cognitive domains. However, a stronger case could be made if the neurons in those areas were observed to demonstrate properties consistent with a general resource function. And this is what has been found. Typically, neurons are selective for specific sorts of information; for example, cells in visual area V1 appear to be orientation specific (Hubel & Wiesel, 1962), whereas cells in V5 appear to be selective for visual motion (Maunsell & Van Essen, 1983). However, invasive electrophysiological recordings in primates have demonstrated that neurons in frontal and parietal association cortex are able to flexibly code for different types of information (Freedman et al., 2001; Toth & Assad, 2002), hinting at their ability to play a role in many different cognitive demands.

But, exactly what general resource function is supported by this network? At present, no clear answer exists, although the literature provides some hints about the boundary conditions regarding its engagement. A recent study demonstrated the importance of attention in activating the frontoparietal network (Hon et al., 2006). In that fMRI study, subjects were presented a series of visual events: some of these they attended to and others they did not. The results of this experiment revealed that only the attended visual events engaged the frontoparietal network. No frontoparietal activity was observed in relation to unattended events, even though these were equivalent to the attended ones.

The idea of general cognitive resources is not new to cognitive psychology (e.g., Broadbent, 1958). For example, concurrent tasks, even when they engage different modalities, are observed to interfere with each other (e.g., Arnell & Duncan, 2002). More tellingly, such interference seems to be modulated by task difficulty. In a concurrent two-task scenario, interference with the second task is greater when the first task requires two responses compared to when it requires only one (Arnell & Duncan, 2002). This suggests that, at some level, the two tasks bear on the same set of resources, and that when performance of the first task “consumes” more of these resources, less is available for the second task.

Outside of the realm of traditional cognitive psychology, the concept of general cognitive resources has also been invoked. In the study of individual differences, cognitive resources are discussed in relation to the construct of intelligence. Although intelligence, as a catch-all category, is no doubt too diffuse an idea to account for all individual differences in cognitive ability (Spearman, 1927), a related construct, general fluid intelligence, is relevant here. General fluid intelligence has been proposed to be involved in reasoning tasks and novel problem-solving ability (Cattell, 1971), and has itself been shown to correlate with performance in a range of cognitive tasks. An intriguing recent study hints at a link between general fluid intelligence and the frontoparietal network we have been discussing. Gray and colleagues (Gray, Chabris, & Braver, 2003) found a positive relationship between frontoparietal activity (elicited when subjects performed a demanding working-memory task) and scores in a traditional test of general fluid intelligence. In the Gray study, subjects who obtained high scores in a traditional test of fluid intelligence activated frontal and parietal areas more than subjects with low fluid intelligence scores when performing a working-memory task.

An issue that arises naturally from the preceding discussion is whether this frontoparietal network can be trained and if so, what the consequences might be. Although these issues have yet to be addressed in great detail, several studies have provided some interesting hints. For example, Rueda and colleagues (Rueda et al., 2005) provided children with training aimed at enhancing executive attention, which has been suggested to be related to our ability to regulate our responses and goal-directed behavior. Other studies have shown that executive attention relies on the frontal and parietal resources previously described (e.g., Fan et al., 2005). In the Rueda study, it was found that 6-year-old children given only five days worth of attentional training showed improvement in tasks tapping attention. However, the more intriguing finding of that study was that the attentional training also appeared to enhance the children's performance on intelligence measures.

What are the implications of this? Although not conclusive, the findings discussed above suggest that there may be real gains in studying the responsiveness of this network to training programs. If general resources are utilized by many different cognitive operations, it may be the case that programs aimed at enhancing the function or efficiency of this network may have widespread positive results, with performance in many different mental tasks being improved or enhanced. Of course, at this point, this idea is largely speculative but it nonetheless warrants investigation.

SPECULATIVE AND CONCLUDING REMARKS

This chapter had, as its aim, the discussion of some concrete examples of how neuroscience techniques are able to provide valuable information to the study of cognitive modification. It is, however, nowhere near an exhaustive review of the many ways that neuroscience can contribute to that endeavor. Although some progress has been made in integrating brain and mind sciences, and benefits have been accrued from this integration, still more work needs to be done.

One obvious area that has not been discussed pertains to the factors that may affect the success of a modification program. The success of such a program is likely to depend on many things. For example, one needs to consider when to introduce a program. Is there a sensitive period for a particular cognitive system that one should target? Additionally, in nature, different people have different levels of cognitive ability. Therefore, one might ask if a given program will be generally effective in the population, or whether or not it needs to be tailored to the recipient. In time to come, neuroscience might be able to provide some answers to these questions. For example, it is most likely that individual differences in cognitive ability are caused by the interaction of genes and the environment. Different genes have been identified that appear to have an influence on behavior. For example, mutations of the FOXP2 gene are associated with compromised linguistic ability, as well as underactivation of Broca's area, an area known to support language processing (Liegeois et al., 2003). Additionally, we also have some knowledge regarding genes that influence parts of well-defined neural circuits. For example, two genes, the dopamine D4 receptor (DRD4) and monoamine oxidase a (MAOA) genes, are known to influence the activation profile of the anterior cingulate, a neural structure known to be involved in cognitive conflict resolution and attention (Fan et al., 2003). What is not known is how genes like these interact with the environment, or with exposure to training programs. A proper understanding of gene-environment interactions is likely to provide invaluable information to researchers or practitioners planning modification programs. For example, knowledge of such interactions may indicate the existence of particular periods during which modification programs will be most effective.

The upshot here is that neuroscience can make valuable contributions to the study and implementation of cognitive modification. It may be that, at the present, these contributions are limited in scope. However, this does not mean that such contributions will remain at this level perpetually. As neuroscience unearths more about how the brain supports the mind, facts may be discovered that might well enhance the effectiveness of attempts at cognitive modification. Consider a situation in which we knew exactly the neural circuit that supported a given cognitive system, as well as the genetic and environmental factors that influence this circuit. In that situation, it may be possible to, using information from a combination of genotyping, brain imaging and behavioral measures, to tailor a modification or enhancement package to a given person, thereby maximizing its efficiency.

References

Ansari, D., & Coch, D. (2006). Bridge over troubled waters: Education and cognitive neuroscience. Trends in Cognitive Sciences, 10, 146–151.

Arnell, K., & Duncan, J. (2002). Separate and shared sources of dual-task cost in stimulus identification and response selection. Cognitive Psychology, 44, 105–147.

Broadbent, D. E. (1958). Perception and Communication. London: Pergamon.

Cabeza, R., & Nyberg, L. (2000). Imaging cognition II: An empirical review of 275 PET and fMRI studies. Journal of Cognitive Neuroscience, 12, 1–47.

Cattell, R. B. (1971). Abilities: Their Structure, Growth and Action. Boston: Houghton-Mifflin.

Chee, M. W., Soon, C. S., Lee, H. L., & Pallier, C. (2004). Left insula activation: A marker for language attainment in bilinguals. Proceedings of the National Academy of Sciences, 101, 15265–15270.

Cohen, L., Lehéricy, S., Chochon, F., Lemer, C., Rivaud, S., & Dehaene, S. (2002). Language-specific tuning of visual cortex? Functional properties of the visual word form area. Brain, 125, 1054–1069.

Dehaene, S., Spelke, E., Pinel, P., Stanescu, R., & Tsivkin, S. (1999). Sources of mathematical thinking: Behavioral and brain-imaging evidence. Science, 284, 970–974.

Duncan, J. (2006). Brain mechanisms of attention. Quarterly Journal of Ex-perimental Psychology, 59, 2–27.

Duncan, J., & Owen, A. M. (2000). Common regions of the human frontal lobe recruited by diverse cognitive demands. Trends in Neuroscience, 23, 475–483.

Fan, J., Fossella, J. A., Summer, T., & Posner, M. I. (2003). Mapping the genetic variation of executive attention onto brain activity. Proceedings of the National Academy of Sciences, 100, 7406–7411.

Fan, J., McCandliss, B. D., Fossella, J., Flombaum, J. I., & Posner, M. I. (2005). The activation of attentional networks. Neuroimage, 26, 471–479.

Fincham, J. M., Carter, C. S., van Veen, V., Stenger, V. A., & Anderson, J. R. (2002). Neural mechanisms of planning: A computational analysis using event-related fMRI. Proceedings of the National Academy of Sciences, 99, 3346–3351.

Freedman, D. J., Riesenhuber, M., Poggio, T., & Miller, E. K. (2001). Categorical representation of visual stimuli in the primate prefrontal cortex. Science, 291, 312–316.

Goswami, U. (2004). Neuroscience & education. British Journal of Educational Psychology, 74, 1–14.

Gray, J. R., Chabris, C. F., & Braver, T. S. (2003). Neural mechanisms of general fluid intelligence. Nature Neuroscience, 6, 316–322.

Hon, N., Epstein, R., Owen, A. M., & Duncan, J. (2006). Frontoparietal activity with minimal decision and control. Journal of Neuroscience, 26, 9805–9809.

Hubel, D. H., & Wiesel, T. N. (1962). Receptive fields, binocular interaction and functional architecture in the cat's visual cortex. Journal of Physiology, 160, 106–154.

Jiang, Y., & Kanwisher, N. (2003). Common neural substrates for response selection across modalities and mapping paradigms. Journal of Cognitive Neuroscience, 15, 1080–1094.

Liegeois, F., Baldeweg, T., Connelly, A., Gadian, D. G., Mishkin, M., & Vargha-Khadem, F. (2003). Language fMRI abnormalities associated with FOXP2 gene mutation. Nature Neuroscience, 6, 1230–1237.

Maunsell, J. R., & Van Essen, D. C. (1983). Functional properties of neurons in middle temporal visual area of the macaque monkey. I. Selectivity for stimulus direction, speed, and orientation. Journal of Neurophysiology, 49, 1127–1147.

Posner, M. I., & Rothbart, M. K. (2005). Influencing brain networks: Implications for education. Trends in Cognitive Sciences, 9, 99–103.

Rueda, M. R., Rothbart, M. K., Saccamanno, L., & Posner, M. I. (2005). Training, maturation and genetic influences on the development of executive attention. Proceedings of the National Academy of Sciences, 102, 14931–14936.

Shaywitz, B. A. et al. (2004). Development of left occipitotemporal systems for skilled reading in children after a phonologically-based intervention. Biological Psychiatry, 55, 926–933.

Spearman, C. (1927). The Abilities of Man. New York: Macmillan.

Stanovich, K. E., Cunningham, A. E., & Cramer, B. (1984). Assessing phonological awareness in kindergarten children: Issues of task comparability. Journal of Experimental Child Psychology, 38, 175–190.

Temple, E., Deutsch, G. K., Poldrack, R. A., Miller, S. L, Tallal, P., Merzenich, M. M., & Gabrieli, J. D. E. (2003). Neural deficits in children with dyslexia ameliorated by behavioral remediation: Evidence from functional MRI. Proceedings of the National Academy of Sciences, 100, 2860–2865.

Toth, L. J., & Assad, J. A. (2002). Dynamic coding of behaviourally relevant stimuli in parietal cortex. Nature, 415, 165–168.

Wagner, A. D., Pare-Blagoev, E. J., Clark, J., & Poldrack, R. A. (2001). Recovering meaning: Left prefrontal cortex guides controlled semantic retrieval. Neuron, 31, 329–338.