Aphasia is an impairment in language function as a result of brain damage (Benson & Ardila, 1996). Impairments include difficulty in expressing and understanding language and also deficits in reading and writing ability. However the impairments of aphasia often go further than just language; adjustment problems, influences from the meaning of language and other psychological consequences are often encountered as a result of language impairment (Benson & Ardila, 1996). Aphasia has several causes; in some cases it develops over time as a consequence of neurodegenerative diseases or brain tumour, however aphasia can also occur suddenly as a result of head injury or most commonly from cerebrovascular accident (i.e. stroke). The middle cerebral artery is the major blood supply for the language areas of the brain and it is this, or the branches of it, which are normally affected by a stroke.
There are many types of aphasia defined by their variations in symptoms. The different forms can be divided into three broad categories; fluent, non-fluent, and pure aphasia (Kolb & Whishaw, 2003). Fluent aphasias include Wernikes aphasia and transcortical sensory aphasia and are generally classified by fluent speech (normal phase length, articulation and prosody) but with an impairment in either auditory verbal comprehension or the repetition of words spoken by others. Non-fluent aphasias include Brocas aphasia and global aphasia, here comprehension is unaffected but speech is non-fluent. Non-fluent denotes either sparse output, decreased phase length, or dysprosody (Benson & Ardila, 1996). Pure aphasias include alexia and agraphia; they are selective impairments in one language function, for example impairments in reading.
In 1961 Paul Broca described a patient who was aphasic as a consequence of a frontal lobe lesion (Selnes & Hillis, 2000). The participant could comprehend speech yet he could only successfully articulate one syllable (‘Tan’). Broca attributed this deficit to the inferior frontal gyrus, and this and the surrounding area is now known as ‘Broca’s area’. In 1963 Broca was the first person to bring attention to the left hemispheres specialisation in language function (Finger, 1994). This was a landmark in aphasia research and provided a strong model which subsequent research would confirm. Carl Wernikes work was one of the many which supported Brocas proposal of the left hemispheres language dominance; however Wernike identified that these left hemisphere lesions can also result in a more sensory form of aphasia. However lesion studies such as that by Broca and Wernike can only progress the investigation of aphasia so far. There is not a one-to-one relationship between lesion location, lesion size and the type of symptoms that it causes, and hence also it is not possible to predict the site or size of a lesion from the symptoms presented. Also lesion studies do not allow us to thoroughly investigate the neural mechanisms underlying recovery.
It is still a valid approach to examine the symptoms of aphasia and correlate the deficits with specific brain lesions; however imaging techniques now allow us to investigate the lesions with a much higher precision, this is one of the many improvements imaging methods have brought to the exploration of aphasia. Imaging techniques have advanced the exploration of aphasia from correlation approaches between brain lesions and aphasic symptoms to a more in depth analysis where we can correlate brain function with language performance. It is possible to investigate both what is damaged and what isn’t working (Khatri & Hier, 2000).
Another benefit imaging techniques have bought to the exploration of aphasia is that they allow us to investigate brain regions associated with language in a normal brain, therefore enabling comparisons to lesions studies and predictions about the nature of the damage. The normal development of language during childhood is one event which has been carefully documented using both behavioural techniques and functional neuroimaging to support theories of language and identify what aspects go wrong during aphasia (Benson & Ardila, 1996).
Through the use of functional imaging we can explore many more questions than was possible when only investigating lesioned brains. Functional imaging also allows us to not only see the location of activation during a brain function but also the relative responsiveness among the regions involved and the connectivity between them. Typically fMRI has been used to identify the all the areas associated with a specific function. Through comparisons with normal activation fMRI can identify areas outside the direct site of damage which contribute to the deficit in language function in aphasic patients (Miura et al, 1999). Positron emission tomography (PET) is one neuroimaging technique which has further confirmed that only acquiring anatomical information is not sufficient to gain an understanding of the brains role in either abnormal or normal language function (Jeffrey, 1987). Most PET research has been conducted using fluorodeoxyglucose (FDG) to measure glucose metablolism. It is widely accepted that changes in glucose uptake reflect changes in tissue activity (Jeffrey, 1987) and therefore FDG can provide an insight into active brain regions in both normal and abnormal brain function. Resting FDG PET studies have demonstrated that cerebral metabolic abnormalities in aphasic stroke patients extend beyond the site of the lesion (Jeffrey, 1987)
Some common findings of PET studies of aphasia are that language function correlates to activity in the posterior temporal and inferior parietal regions, and different patterns of glucose asymmetry between homologous areas of the two hemispheres were found in prefrontal regions and each pattern can be associated to different aphasic categories. However temporoparietal glucose metabolic asymmetry is relatively consistent across all aphasic conditions (Jeffrey, 1987).
To further understand how the brain performs language it is important to understand how the regions involved interact and process language over time. fMRI and PET have low temporal resolution and so cannot present us with an accurate temporal scale in which language function is performed. However another two techniques, electroencephalography (EEG) and magnetoencephalography (MEG) have much higher temporal resolution and can map language processes as they unfold in real time (milliseconds) (Friederici, 2001). (add eeg stuff here)
Ground-breaking research into language was performed in 1954 when neurosurgeons Wilder Penfield and George Ojemann used direct cortical stimulation to probe language functions in awake patients (Penfield & Jasper, 1954). They found that direct current to a specific brain region could disrupt linguistic processes or even cause spontaneous vocalisations (Devlin & Watkins, 2007). Work from direct cortical stimulation still continues to provide insights into language function but it is sparsely used due to its highly invasive nature. As an alternative transcranial magnetic stimulation is a non-invasive technique which provides similar stimuluation as direct cortical stimulation to a cortical region with the intention of disrupting information processing. Transcranial magnetic stimulation is a major technique for investigating language function. Through this process we can also expand upon functional neuroimaging methods by exploring causal relationships between brain regions and language function and we can do this with high spatial and temporal resolution (Devlin & Watkins, 2007).
Patient studies have limitations which include differences in premorbid ability, compensatory plasticity, and large variety of lesion site and sizes. TMS avoids these issues through producing a focal ‘virtual lesion’ for too brief a period for compensatory mechanisms to be formed.
Speech-arrest was one of the first outcomes of a language based TMS study. Pascual-Leone et al (1991, as cited by Devlin & Watkins, 2007) invested the neural basis of speech production and induced speech arrest in patients after stimulation of the left inferior frontal cortex. They also recorded that no speech arrest was found after any right hemisphere stimulation, indicating a left hemisphere language dominance. Expanding upon these findings Aziz-Zadeh et al, (2005) found that speech arrest can be induced by stimulation of two specific sites within the left inferior frontal cortex, with the more prosterior site produced speech arrest in both hemispheres and also produced a facial muscle response, whereas the more anterior site only produced speech arrest when stimulated on the left hemisphere and this did not produce a facial muscle response. Aziz-Zadeh et al, (2005) conclude that the more posterior site are connected to motor and premotor regions and consequently stimulating this region in either hemisphere would produce speech arrest via disrupting the motor output of speech, and that the anterior region of the left inferior frontal cortex may be associated with language areas of the prefrontal cortex such as Broca’s area and so stimulation here interferes with the formation of an articulatory plan. However further work is required to confirm this hypothesis.
More specific language processes have been investigated with TMS for example the involvement of Broca’s area in syntactic processing was explored by Sakai et al, (2002). They applied TMS to Broca’s area at varying timepoints during a task in which participants had to identify sentences as correct, grammatically incorrect or semantically incorrect. They found that TMS facilitated the participants reaction times for syntactic but not semantic decisions and concluded that Broca’s area is causally involved in syntactic processing. This provides evidence for the involvement of Broca’s area in Broca’s aphasia in which syntax is affected.
In addition to its research advantages for investigating language functions TMS can also be used with neurological patients to investigate the mechanisms of recovery and has also been used to enhance recovery in some patients (Devlin & Watkins, 2007).
It has been suggested that after damage to the left hemisphere the recovery of language functions such as speech comprehension may depend on a mechanism in the right hemisphere (Crinon & Price, 2005). TMS has been particularly beneficial in investigating the role of the right hemisphere in the recovery process of aphasia patients. Coslett & Monsul (1994, as cited by Devlin & Watkins, 2007) applied TMS to the right temporo-parietal area of patients and controls during a reading task. In patients a single TMS pulse significantly impaired their reading ability, however no affect was recorded from control trials. This finding provides support for the theory that the right hemisphere has increased importance for language function after left hemisphere damage and therefore is important in language recovery. Crinon and Price (2005) support this theory as using fMRI they found that in patients with damage to Wernike’s area language comprehension ability was positively correlated with activation in the right lateral superior temporal region.
As a result of stroke the brain is often imflamed and swollen, as this goes down during the first few days post stroke language function can partially recover. The literature suggests that the severity of the stroke seems to determine the timeline of recovery; with the brain able to recover faster from milder strokes than more sever strokes. However this timeline is poorly understood and some patients can continue to recover months and years post stroke. Saur et al (2006) attempted to investigate the recovery timeline through repeated fMRI examinations paralleled with language testing. Their results suggest that language recovery involves three phases of brain reorganisation; acture, sub-acute, and chronic phases. The acute phase occurs approximately 1.8 days following damage and during this phase activity in the remaining intact left hemisphere language areas is strongly reduced. Approximatlly 12.1 days post-stroke is the sub-acute phase where there is an upregulation of the homologous right language areas which correlates with language improvement. In the chronic phase (approximately 321 days post-stroke) normalisation of the left language areas occur as they regain peak activation and this is correlated with further language improvements. This research has implications for the strategies applied in treating aphasia; what areas to target treatment at and when. Saur et al (2006) propose that both the right and the left hemisphere support recovery, however this fact remains unclear.
Like lesion studies, but with increased accuracy, the combination of perfusion-weighted imaging (PWI) and diffusion-weighted imaging (DWI) can reveal areas of tissue and vasculature damage, and this combined with cognitive assessment can reveal the specific brain regions which are essential for a particular task (Hillis et al, 2006). Hillis et al (2006) provided additional evidence for the function of certain brain regions by not only deeming them essential for a task, through linking impaired regions to impaired functioning, but also showing that recovery of tissue in the damaged areas resulted in an improvement of their associated function. Using hypotheses generated from functional imaging studies about the language networks within the brain areas they restored blood flow to language areas affected by stroke. Consequently they identified the areas of the brain where this blood flow restoration resulted in an improvement of picture naming. The areas which this novel methodology identified as essential were the left posterior middle temporal/fusiform gyrus, Broca’s area, and Wernicke’s areas. Therefore their results show that these areas were associated with an improvement in picture naming and provide evidence for the left hemispheres involvement in recovery through improved tissue function rather than reorganisation of structure-function relationships.
Some studies have shown that TMS can be used to aid the recovery process in aphasia patients; this is most often when TMS is repeatedly applied over many days or weeks and long term benefits have been reported. Repeated TMS (rTMS) is has been used for rehabilitation of aphasic patients and has works by decreasing excitability in the targeted cortical area which lasts beyond the duration of direct stimulation therefore leading to measurable behavioural effects. Such effects have been found to be beneficial in a number of experiments, including those where regions of the right hemisphere have been stimulated to prevent any inhibitory processes that it may be performing and therefore allowing reactivation of some language areas within the left hemisphere which facilitates functional recovery (Lefaucheur et al, 2006). Martin et al, (2009) also support these findings that TMS can improve behaviour as they reported long term improvements of language function after rTMS on patients with chronic non fluent aphasia. Patients with non fluent aphasia often display increased activity in the right hemisphere with no behavioural changes, this suggest that this activity may be maladaptive. Naeser et al (2005, as cited by Devlin & Watkins, 2007)reported that when rTMS was applied to the rostral right inferior frontal gyrus patients improved at picture naming. After a 10 day regime of 20 minutes rTMS a day facilitatory effects of naming were still found 8 months later. However such studies are based on a small groups of patients and very few have control groups, therefore these results need to be tested on a much bigger scale to determine if this technique should become an established rehabilitative therapy to improve recovery in certain aphasic patients (Martin et al, 2009).
The combination of TMS and other imaging modalities is essential to obtain an accurate understanding recovery and the benefits TMS can provide, for example functional imaging can be used before and after TMS to explore the plasticity of the neural networks underlying language function (Martin et al, 2009).
Around eighty thousand new cases of aphasia occur annually (Khatri & Hier, 2000), consequently prognostic indicators of this condition will make a vital difference. There currently are no reliable ways to predict individual outcomes of the condition.
It would be beneficial to understand and predict the pathway of recovery. A better understanding of recovery will lead to more beneficial clinical interventions (Khatri & Hier, 2000). Recently one way in which this is currently being effectively explored is through computer modelling. Through inputting empirical information about anatomical and functional impairments a computer model of aphasia can be simulated, and if data about relearning during recovery is also provided the model can update and provide a model of relearning after lesion (Welbourne & Lambon Ralph, 2006). The more data provided the more accurate these models will be.
The role of the right hemisphere needs further exploration as opinions on its role are mixed. There are recorded cases of patients with language impairment as a result of a left hemisphere stroke, whose language was further impaired by a second stroke in the right hemisphere (Basso et al, 1996). Such cases highlight that the right hemisphere must also contribute to language processing (Crinon & Price, 2005). Some research shows that the right hemisphere is involved in recovery of language function, for example new activity in the right hemisphere has been found as a result of speech therapy in aphasic patients (Raboyeau et al, 2008). However contrasting this hypothesis are studies which suggest increased right hemispheric activity detrimentally interferes with performance on language tasks (Devlin & Watkins, 2007) and may even hinder recovery of language such as Saur et al (2006) who reported that increased activity in the right IFG was not associated with improved naming ability. Pre-morbid language organisation is highly likely to influence the results from these studies and in time it should be possible to determine with functional imaging those patients in which right hemisphere activity is detrimental to language function and those in which it is beneficial. Clarification of the role of the right hemisphere is needed as the results may have strong implications for the development of treatments of aphasia. Intact brain regions are a target for rehabilitative techniques and therefore following left hemispheric damage areas in the right hemisphere may be an advantageous target for therapy.
There are many improvements and clarifications that need to be made within the study of aphasia. Neuroimaging techniques have become important tools for studying language at both the cognitive and neural levels, and further technological and methodological advances will create even greater opportunities for language research.
Research Needs in Neuroimaging Related Aphasia Rehabilitation/Treatment:
Utilize neuroimaging techniques to document the neurobiological correlates of treatment effects
Utilize imaging techniques to document the effect of particular treatment on brain functioning and/or to predict functional outcome/likelihood of retention of skills post-treatment
Determine utility of using neuroimaging data to determine the potential benefit of particular treatments
Examine the effects of timing of treatment on neural recruitment patterns, ie, early after a deficit, or after a delay (when acute changes have at least partially dissipated).
Study the effectiveness of components of intervention as well as combinations of intervention (for example, behavioral and pharmacotherapeutic);
Examine features of training: presence of feedback associated with each trial, context of stimulus use, the timing of stimulation, what the individual already has mastered in area to be trained
Identify the “cost” of treatment, ie, the possible detrimental effects of treatment, on other (nontreated) cognitive/language processes. Does competence in another domain suffer as treatment results in improvement in targeted area?
Develop approaches to documenting changes over time within the intervention setting
Determine the reliability and stability of measurement
Develop reliable, quantitative outcome measures
Determine efficacy of treatment as a function of intensity of treatment (focused, intense intervention vs. more traditional, less intense treatment) and the relation between intensity of treatment and recruitment of neural tissue in recovery
Identify responses which reflect restoration versus reorganization of function, that is, training to the deficit vs. teaching of compensatory strategies
Identify strategies that facilitate generalization of behaviors learned in treatment
Develop more effective intervention/therapy tools with detailed specification of training procedures, task difficulty/patient performance
Determine specificity of training effects.
Determine role of general regulatory factors (attention, motivation, mood, level of arousal) in efficacy of treatment.