Sophie Scott, 9^th Feb, 2000

Boston Aphasia Classification System

1. Broca’s aphasia: a lesion involving the expressive speech centre

2. Wernicke’s aphasia is caused by a lesion of the audio-verbal centre

3. Conduction aphasia results from a lesion of the pathways connecting the audio-verbal and expressive speech areas.

4. Global Aphasia is produced by an extensive lesion involving both the audio-verbal centre and the expressive speech centre.

5. Transcortical motor aphasia is associated with disruption of the pathways connecting the concept area to the expressive speech area.

6. Transcortical sensory aphasia results from lesions of the pathways connecting the audio-verbal centre to the concept centre.

7. Isolation aphasia is caused by lesions, which disconnect the concept centre from the audio-verbal centre and the expressive speech centre.

8. Anomic aphasia is produced by a lesion involving the pathways which connect the concept centre to the expressive speech centre, or if comprehension is also disrupted, a lesion of the concept centre.

Non fluent speech, poor repetition, spared comprehension. Long pauses in effortful speech, some repetition, often also a motor speech problem (apraxia) or dysarthria accompanies this.

Normally caused by damage to the lateral portion of the primary motor strip, so often hemiplegic on contralesional side. Show some pathological reflexes on this side (e.g. Babinski). Ideomotor apraxia (carry out commands that can be done spontaneously) in the ‘good’ hand.

Is all the damage in the same place?

Classical Broca’s area is BA 44/45 – posterior third of the inferior frontal gyrus. Mohr et al (1978) looked at 20 lesions at autopsy. A lesion to Broca’s area itself caused transient speech apraxia. Transient, because the right hemisphere probably takes over. For full "Broca’s aphasia" the lesion needed to encompass a lot of the Sylvian region incl. Left opercula, insula, and subjacent white matter in the territory of the middle cerebral artery inc. Broca’s area. Broca’s original case involved these areas. It is worth noting that architectonically, BA 44 is like premotor cortex, and BA 45 is like prefrontal cortex, so it is unlikely that they perform the same processing functions.

A disorder of comprehension, also poor naming and repetition. Speech output is fluent, with lots of paraphasia’s and monitoring problems. Don’t normally have a hemiparesis, may have a visual field defect (e.g. a hemianopia). Classically Wernicke’s aphasia arises from a left tempero-parietal lesion

Basso et al (1995) 8 cases of Wernicke’s aphasia, extensive perisylvian lesions. The role of the sub cortical regions is not clear.

Post-rolandic lesions are associated with fluent aphasias.

1. If make lexical errors - lesion associated with infero-lateral temporal coretx (middle and inferior temporal gyri), also parietal cortex.
2. Phonological errors - lesions of Wernicke’s area proper - posterior STG and supramarginal gurus.

Disruption of repetition, with preserved (relatively) comprehension and spontaneous speech, though there may be phonemic errors. Some problems with confrontation naming. There tends to be no associated neurological problem, e.g. hemiparesis is rare; if there is weakness, it is the arm more than the leg. Some patients have a limited right hemianasthesia, or visual field defect. May have difficulty with bucco-facial or limb movements to command.

This is typically taken to be a disruption of the actuate fasiculus (Wernicke, Geshwind). It is known as a disconnection syndrome because of this. The actuate fasiculus (AF) is a white matter tract running from the posterior STG to Broca’s area. The lesions involved run from beneath Wernicke’s area and the supramarginal gyrus so it is the posterior AF that is involved. However other reports have implicated the cortex in this condition, esp. in Wernicke’s area (Kleist, 1962, Mendez and Benson 1985). This latter case involved 3 cases without AF lesions, 2 in the left temporal-parietal regions and on on the right (such as case, in a right handed patient, is called a crossed aphasia). More recent electrical stimulation studies have backed this up. For example, Anderson et al (1999) worked with an epileptic patient who, when her posterior STG was stimulated, made phonemic paraphasic errors and had impaired repetition, with preserved semantic comprehension. Neuroanatomical work with humans (which is very rare) has shown that the main projections from primary auditory cortex run anteriorly down the STG. The posterior STG and STS do not have direct anatomical connections from anterior STG. Thus, there may be dissociation of function between anterior and posterior STG/STS, and this may underlie the difference between lesions in these areas. Which brings us to:

All major language functions are impaired, output and comprehension. Tends to follow an extensive left hemisphere lesion involving Broca’s are and Wernicke’s area, such large scale brain damage involves many associated neurological signs: hemiplegia, sensory loss, visual field defects and attentional disturbances such as extinction or neglect.

The lesions need not be large; they can often spare Wernicke’s area. Involves both cortical and subcortical regions. This sparing of Wernicke’s area with comprehension problems may well be a result of the different pathways for input in the auditory system mentioned before. Anterior STG may be more important in comprehension.

Repetition OK – comprehension and spontaneous speech compromised, the latter strikingly so. The repetitions are not mandatory (i.e. no echolalia). Can correct errors in what they are asked to repeat. Spontaneous speech is stumbling, and stuttering, agrammatic and simple. Can count. Similar neurological signs to Broca’s aphasia. Sensory loss and visual field problems are not characteristic. Lesions are not well understood, but seem to be anterior and superior to Broca’s area.

Impaired comprehension, preserved repetition and fluent output. Often include words heard in their output without understanding them. Such ‘repetition’ is mandatory, i.e. the patients are echolalic. Tend not to show a basic neurological deficit, though may have a sensory deficit. Kertesz et al (1982) carried out a study into the localisation of the lesions, and found 2 sites; medial inferior ventral temporal lobe, and anterior STG.

Only repetition – comprehension and spontaneous speech are compromised. No voluntary language use. Variable neurological signs. Some have a bilateral paralysis, leading to quadriplegia or quadriparesis, or unilateral signs, such as a right hemiplegia. There is often a sensory loss. Two post-mortem studies indicated that anterior and posterior damage was involved (Geschwind et al, 1968, Whitaker 1976). A CT scan by Chenery and Murdoch (1986) showed no focal lesions. A study by Ross (1980) showed left motor and sensory cortices, also parietal lobe involvement.

Word finding difficulties are the most prominent feature, leading to ‘vague’ speech. Most cases have no associated neurological signs. Is regarded as non-localising as no one area is involved from patient studies. Gloning et al (1963) found that 60% of their patients had temporal-parietal lesions, but the other 40% were wide ranging, though all in the left hemisphere.

Controversial, and has arisen since we have had good brain imaging techniques. May arise due to associated problems of the damage, e.g. a large haemorrhage into subcortical regions can result in pressure, pressing the cortex against the skull and leading to problems distal to the infarct. Ischaemic infarcts in the thalamus itself do lead to language problems, though the clinical picture is not always clear, but mostly like a transcortical aphasia, with preserved repetition and variable comprehension, but a problem with spontaneous speech. The outcome is variable.

Striato-capsular lesions – there my be an anterior-posterior dimension, e.g. Naeser et al showed that anterior lesions, including capsular-putamen lesions have good comprehension, slow grammatical speech, dysarthria and hemiplegia. More posterior infarcts showed poor comprehension, fluent Wernicke type speech and right hemiplegia. Both anterior and posterior white matter extensions – global aphasia. These cases did involve cortical regions as well, and Naeser (1997) has now made the claim that sub cortcial aphasia does not exist. Again, the presentation is varied, as is the outcome.

The key route for basal ganglia involvement in speech and language is in the motor loops. Some have made the claim that the thalamus mediates between language areas (Penfield and Roberts, 1956). Others have claimed that the thalamus controls and releases pre-formed speech acts, others that this is the role of the basal ganglia. Interesting point with respect to the utterances and movements of people with Tourettes.

An early PET study, involving 5 different conditions:

1. rest
2. perception of non words
3. noun noun comparisons (fruit – apple…..furniture-shirt)
4. verb noun comparisons (e.g. eat – apple….knit-spectacles
5. verb generation – similar to the verbal fluency test, presented with a noun (e.g. flower) and had to generate as many associated verbs as possible.

The stimuli were presented over headphones. Subjects indicated the responses to (3) and (4) with their fingers. They made the same response when they thought of a verb in (5). During rest and non-words the subjects were asked to make occasional finger movements. All the nouns were concrete.

Rationale:

2 minus 1 = prelexical phonetic processing

3 minus 2 = post phonetic semantic processing of nouns

4 minus 2 = post phonetic semantic processing of verbs

5 minus 4 = regions involved in the retrieval of verbs from memory

The subjects were silent in all conditions, to avoid activation due to vocalisation.

Results

1. Listening to non-words – activated bilateral primary and secondary auditory cortex, i.e. along both superior temporal gyri.
2. Noun-noun and noun-verb judgements activated similar regions to non word perception, when compared to rest, and this was confirmed when they were compared directly against non word perception. Direct comparison of one against the other revealed no verb – noun difference.
3. In the verb generation condition, there was no activity in auditory cortex and little in left STG, since the rate of heard words was very low. Posterior STG (‘Wernicke’s area’) was active. In more anterior regions, there was activity in this condition in the posterior left inferior frontal gyrus (Broca’s area), posterior left middle frontal gyrus and the supplementary motor area. If left SMA is lesioned, speech output is lost although comprehension is preserved. This activation may reflect inner speech. Posterior middle frontal gyrus is associated with transcortical motor aphasia, and is involved when several verbs must be generated for each noun, rather like spontaneous speech. Silent word retrieval involves 2 areas in left dorso-lateral prefrontal cortex, left posterior STG, and the SMA.

This shows that in the ‘normal’ brain, there is a close correspondence between the brain regions involved in the deliberate, voluntary generation of words, and the brain regions involved in Broca’s aphasia and transcortical motor aphasia. There is a complex network of regions subserving this – and that without overt articulation.

In the above study, speech perception led to extensive bilateral activation in the dorso-lateral temporal cortex, compared to silent rest. However heard speech is an acoustic signal, so need to control for ‘hearing a sound’ activation in primary auditory cortex. In addition, speech is a complex acoustic signal, so need to control for some aspects of acoustic complexity, without making a sound that is too speech like, so as to avoid invoking a ‘speech mode’ or perception. Also wanted to avoid using a subtractive design.

1. Non speech stimuli: The non-speech stimuli used here was signal correlated noise (SCN; Schroeder, 1968) which forms a stimulus with the same instantaneous amplitude as the speech signal, but with a white spectrum; it thus contains none of the spectral information such as transients and formants that lead to intelligible speech. SCN preserves the amplitude envelope of speech sounds, so these stimuli have preserved segmental information about manner of articulation, and the tempo, rhythm and syllabicity of the speech without comprehension being possible (Rosen, 1992); it thus forms an appropriately complex non-speech contrast.
2. The design was parametric – the rate of presentation of speech and SCN was varied across scans. Over the 12 scans, the subjects heard either There were twelve PET scanning conditions (six presentation rates and two experimental conditions - speech and SCN), which were randomized within and across subjects to control for time effects. During scanning, the normal subjects heard either spoken words or SCN sounds, presented binaurally, at rates of 1, 5, 15, 30, 50 and 75 per min. In the 1 per min condition subjects heard one stimulus before the scan started and they anticipated hearing further stimuli. The presented words were bisyllabic nouns, matched for frequency concreteness and imageability. There were 6 subjects, all right-handed males.

Could look at brain regions that increased in rate with both increasing rates of speech and SCN. This showed bilateral primary auditory cortex and STG lateral to this.

Could also look for brain regions that increased in rate with speech only. This showed bilateral STS, both anterior and ventral to the primary auditory cortex. There was increasing activity with increasing numbers of speech stimuli in these more anterior regions. There was one region of asymmetry - a response in the left posterior ventral STG/STS, i.e. in the core of Wernicke’s area.

Thus there is a bilateral involvement of STG/STG in speech perception, though speech and SCN are still different, this implicated both hemispheres – recovery? Word deafness?

Also – Wernicke’s area is asymmetric – different function?

1. Repetition of single bisyllabic nouns at 10, 20, 30, 40, 50 per minute
2. Listening pairs of at 20, 40, 60, 80, 100 per minute
3. Anticipation (of listening)

This enabled us to hold constant the number of words heard, and contrast:

Listening with anticipation – as before this shows bilateral STG

Repetition minus anticipation. This shows the motor system – right and left primary sensori-motor cortex, and left anterior insula, dorsal brainstem, left and right rostral cerebellum, as well as bilateral STG

Repetition minus Listening (just the motor act), as above, plus left posterior pallidum and the anterior cingulate gyrus.

Listening minus repetition: a little right primary auditory cortex, and a left anterior STG.

When repeating, the bilateral sensorimotor activation corresponds to controlling the muscles used in speech. Dysarthria can arise from cerebellar lesions, as corresponding to the involvement in repeating. The brainstem activation may arise from the involvement of the trigeminal and facial nuclei. Left anterior insula is involved in articulation. The involvement of the left opercular cortex in transient speech arrest seen by Mohr et al may have been due to an effect of disruption without permanent damage on the immediately adjacent insular cortex. Dronkers (1996) implicates the left anterior insula in speech apraxia – a disorder of shaping the vocal tract for a particular speech sound (spatial co-ordination disrupted) and to reshape for the next sound (temporal co-ordination disrupted). The pallidum is involed in basal motor loops between motor and premotor cortex, striatum, pallidum, thalamus, brainstem and cerebellum. Bilateral pallidal disease results in severe dysarthria; the left dominance here probably reflects the left hemisphere dominance for language. There was a deactivation in the left mid STG/STS when listening to own voice – suggesting an attentional role? More important to listen to incoming stimuli when repeating?

Wise R., Chollet F., Hadar U., Friston K., Hoffner E., Frackowiak R. (1991) "Distribution of cortical neural networks involved in word comprehension and word retrieval". Brain 114, 1803-1817.

Wise, R.J.S., Greene, J., Büchel, C., Scott, S. K. (1999) Brain systems for word perception and articulation The Lancet, 353 (9158), 1057-1061.

Mummery, C.JAshburner, J., ., Scott, S. K., and Wise, R.J.S. (1999) Functional neuroimaging of speech perception in six normal and two aphasic patients. Journal of the Acoustical Society of America, 106 (1) 449-457.