3. Neurological processing of hierarchically organised sequences in non-linguistic domains
A broader perspective sees grammar as just one of many hierarchically organised behaviours being processed in similar, prefrontal neurological regions (Greenfield, 1991; Givon, 1998). As Broca’s area is found to be functionally salient in grammatical processing, it is logical to assume that this is the place to search for activity in analogous hierarchical sequences. Such is the basis for studies into music (Maess et al., 2001), action planning (Koechlin and Jubault, 2006) and tool-production (Stout et al., 2008).
3.1 Music and Broca’s Area
Music and language are often cited as cohorts in their use of PFC processing (Greenfield, 1991). Earlier studies (Givon, 1998) show those with impaired language skills also have impaired rhythmic-musical abilities. After all, music performance “[…] is indeed a complex rhythmic-hierarchic skill par excellence. It is perhaps the closest analog – thus potentially dependent on a neural homolog – of the syntactic structure of language.” (Givon, 1998, pg. 154).
It seems these previous predications are somewhat vindicated by a few related studies into musical syntax (cf. Patel, 2003). In one experiment performed by Maess et al. (2001), participants (specifically non-musicians) listen to sequences of five harmonically related chords while scanned by MEG (magnetoencephalography). In some sequences, different chords (Neapolitan sixth) are placed in the third and fifth harmonic positions. According to the study, the Neapolitan sixth chords elicit the most dissonance when in the fifth position, which is where they found neural activity in Broca’s area and its right hemisphere homologue – especially BA 44.
Despite the findings of such studies, which indicate a pooling of resources in processing both language and music syntaxes, there is not to my knowledge anything to negate the charge of musical ability merely being a secondary function of language (Patel, 2003). This severely limits the conclusions one can infer about domain-specificity of Broca’s area. However, these studies are not in isolation and we can seek some confirmation from other, more disassociated domains.
3.2 Action Plans and Broca’s Area
According to Koechlin and Jubualt (2006), several previous studies show “the temporal dimension of executive control is processed in the lateral prefrontal cortex by a top-down control system… extending from premotor to the most anterior prefrontal regions.” (pg. 963). In this top-down control system of action planning both hierarchical and temporal forms of organisation are taking place, with each of these executive[1] processes mirroring the differences found in sentence complexity. But instead of (AB)n and AnBn rules, the authors use simple action chunks (temporal) and superordinate chunks (hierarchical).
In testing for neural activation in the temporal tasks, participants perform pre-learned sequences of button presses in response to a particular stimulus: choosing between left, right or both left and right buttons. For the supordinate condition, subjects perform pre-learnt categorisation sequences, each of which corresponds to a different stimulus. In response to each stimulus, “[…] subjects pressed the left or right button appropriate to the current categorization task before inferring the next categorization task of the learned sequence in order to respond correctly to the next stimulus.” (ibid, pg. 964).
From these sequencing tasks, Koechlin and Jubault find bilateral activation of Broca’s area and its right homologue. Of greater relevance is the disparity found: simple action chunks correspond with BA 44, while the slightly anterior BA 45 activates in response to superordinate chunks. This apparent gradient of activation fits in with wider suggestions for the role of the PFC (Hagoort, 2005), where processing takes place across an anterior-posterior axis (Koechlin and Summerfield, 2007).
Their findings also offer an explanation as to why activation of Broca’s area corresponds to other areas such as working memory, which: “[…] involves rehearsal and hierarchical reorganization of mental representations of action in memory. These behaviors are based on chunking action into nested functional segments, including simple and superordinate chunk-like sequences of motor acts, sensorimotor mappings, or sequences of sensorimotor mappings.” (Koechlin and Jubault, 2006, pg. 971).
Thus, activation during the processing of syntactic movement may not reflect working memory load in Broca’s area, as suggested by Santi and Grodzinksy (2007). Instead, memory information might be stored in the posterior cortical association areas, which are then influenced when Broca’s area is performing a chunking operation (Curtis and D’Esposito, 2003). Bor et al. (2003) offer further support for this, showing that by strategically encoding information frontal cortical activity increases, even though the working memory load has decreased.
3.3 Tool Manufacture and Broca’s Area
Linking tool-use and language is nothing new. Charles Darwin (1871) and, slightly later, Frederick Engels (1883) see tools and language sharing an intimate evolutionary link, with latter going as far as to say these aspects are the “[…] two most essential stimuli under the influence of which the brain of the ape gradually changed into that of man.” (Engels, 1883, as cited in Stout, 2008b).
Like previous non-linguistic behaviours discussed, the use and manufacture of tools also displays evidence of hierarchical complexity. However, unlike music and action plans the neurological role is slightly more ambiguous. For instance, a recent study found the use of tools shows “[…] no evidence to suggest that Broca’s area is a critical substrate for representing these skills once established, at least not in the adult modern human brain.” (Scott-Frey, 2008, pg. 1954). Instead, there is apparently a more distributed overlap between tool use and language circuits in relation to communicative gestures.
Still, the story of tools does not end with just their usage. Manufacturing tools involves a related but quite different skill set. Bearing this in mind, Stout et al. (2008) devise a method to test how different levels of tool complexity are represented in the brains. Specifically, the study investigates tool-making abilities of modern humans in two distinct Early Stone Age (ESA) industries. First are the Oldowan tools, which first appear in the fossil record around 2.6 million years ago, and consist of stone chips made from a bigger core stone through repeated strikes from a hammerstone (ibid). Appearing later on in the fossil record (~250,000 years ago) are Acheulean-based tools. These demonstrate a greater degree of specialisation, such as the Acheulean hand axe, which:
“[…] seems much more demanding than Oldowan flaking, requiring greater motor skill and practical understanding of stone fracture…more elaborate planning of immediate goals to long-term objectives… [and] an increased number of special purpose knapping tools and technical operations.” (ibid, pg. 1941).
This improvement in technology also coincides with increased brain encephalization in the hominin lineage, and suggests the Acheulean technology might show different activation patterns to that of Oldowan tools. Building on previous work (Stout and Chaminade, 2007) looking at novice Oldowan tool makers, Stout and his colleagues perform PET scans on expert artisans while they make their ESA tools. From their results, it seems that Oldowan technology utilises posterior, sensorimotor portions of the brain[2]. Making Acheulean hand axes, however, shows increased activation in the phylogenetically younger PFC – in particular, the right hemisphere homologue of Broca’s area (BA 45) (ibid).
Considering the activations in tool-manufacturing, it appears the overarching emphasis found in all these non-linguistic domains is how temporally organised processes are automated; becoming subsets of hierarchically structured behavioural plans (Koechlin and Summerfield, 2007; Stout et al., 2007). This notion is supported by Fiebach and Schubotz (2006) who, in reference to grammar, add: “[…] the complexity of rule-based processes carried out in the PMv/Broca region varies along a posterior-to-anterior gradient…More anterior regions (BA 44 and 44/45) utilize these rules to adjust syntactic predications when processing complex sentences and to relate grammatically more complex sentence structures back to the simplicity-based structural templates” (pg. 501).
4. Conclusion
It seems intuitively obvious that aspects considered to be uniquely human entail similar cognitive demands. Thus, the processing of hierarchical structures may be one of the key roles Broca’s area subserves. By extension, elements of language, such as syntactical complexity, are just one of many complex behaviours sharing a common neurological substrate. But let us not be too hasty in arriving at these conclusions. In attempting to integrate these various forms of human behaviour, attention must be paid to several other factors. For instance, there are still valid objections to Broca’s area being a domain-general region (cf. Grodzinsky and Santi, 2008). Furthermore, even if the domain-general account is accurate, assigning it an overarching function may be detrimental in enhancing our understanding.
That these differing interpretations exist does not hinder the argument put forward in this essay. Instead, it simply provides an impetus for more research to help clarify issues and avoid generalisations. Broca’s area is probably not delimited to one function, but at the same time it is certainly not an all-encompassing explanation for human behaviours. Also, there are clear cytoarchitectural and functional difference between BA 44 and BA 45, which will need to be clearly defined in the context of behavioural organisation. However, given the literature discussed, a common denominator across these linguistic and non-linguistic areas points towards Broca’s area serving as a domain-general, neurological substrate involved in processing hierarchically organised sequences.
Click here for the first part.
[1] Essentially a conscious, goal-oriented behaviour.
[2] Slightly unrelated to the essay at hand, but the study also showed neurological differences between novice and skilled Oldowan tool-makers. The latter demonstrates a more distributed, bilateral activation. Meanwhile, novice tool-makers show a left-lateralised preference. Stout et al. suggest this cross modality activation might be related to the distributed network found in language; however, this is mere speculation.
Main References
Maess B, Koelsch S, Gunter TC, & Friederici AD (2001). Musical syntax is processed in Broca’s area: an MEG study. Nature neuroscience, 4 (5), 540-5 PMID: 11319564
Koechlin E, & Jubault T (2006). Broca’s area and the hierarchical organization of human behavior. Neuron, 50 (6), 963-74 PMID: 16772176
Stout D, Toth N, Schick K, & Chaminade T (2008). Neural correlates of Early Stone Age toolmaking: technology, language and cognition in human evolution. Philosophical transactions of the Royal Society of London. Series B, Biological sciences, 363 (1499), 1939-49 PMID: 18292067
Your site was extremely interesting, especially since I was searching for thoughts on this subject last Thursday.