Acquired amusia after a right middle cerebral artery infarction – a case study

ABSTRACT

A 62-year-old musician—MM—developed amusia after a right middle-cerebral-artery infarction. Initially, MM showed melodic deficits while discriminating pitch-related differences in melodies, musical memory problems, and impaired sensitivity to tonal structures, but normal pitch discrimination and spectral resolution thresholds, and normal cognitive and language abilities. His rhythmic processing was intact when pitch variations were removed. After 3 months, MM showed a large improvement in his sensitivity to tonality, but persistent melodic deficits and a decline in perceiving the metric structure of rhythmic sequences. We also found visual cues aided melodic processing, which is novel and beneficial for future rehabilitation practice.

KEYWORDS:

• Acquired amusia
• pitch processing
• rhythm
• memory
• visual aids

Introduction

Amusia is a musical disorder characterized by impairments of musical perception and/or musical production (Peretz et al., 2002). Congenital amusia refers to amusia associated with abnormal development of musical skills, whereas acquired amusia is caused by brain injury such as a stroke. Cases of acquired amusia provide a unique opportunity to understand the relation between lesioned neural structures and the associated musical impairments. Identifying the neurological correlates of acquired amusia may also inform rehabilitation strategies.

Acquired amusia can impact various aspects of musical processing, including the ability to discriminate between melodies that differ in contour or melodic intervals, rhythm discrimination, sensitivity to functional relationships between notes (i.e., tonality), and/or musical memory (Ayotte et al., 2000; Schuppert et al., 2000; Steinke et al., 2001; Stewart et al., 2006; Zatorre & Samson, 1991). These impairments can co-occur or can be dissociated with each other depending upon the location of lesions. For example, Sihvonen et al. (2016) found both pitch and rhythm deficits in acquired amusia were associated with an extensive lesion overlap in superior temporal gyrus (STG), Heschl’s gyrus and the right striatum, though lesions in the right dorsal striatum were the most significant for rhythm deficits. In contrast, Zatorre and Samson (1991) reported that people with focal excision in the right temporal lobe showed a significant deficit in retention of tonal information but not in other forms of pitch discrimination.

People can recover from acquired amusia in 3–6 months after a stroke but do not always do so. Sihvonen and colleagues found that the poor recovery from acquired amusia was associated with the atrophy in areas adjacent to the initial lesion and the damage to the right frontotemporal regions, whereas the good recovery was associated with increased activation and functional connectivity in bilateral frontoparietal regions (Sihvonen et al., 2016; Sihvonen, Ripollés, et al., 2017; Sihvonen, Särkämö, et al., 2017). More specifically, poor recovery was linked to right anterior STG/middle temporal gyrus (MTG) and inferior parietal lobule (IPL) atrophy in pitch amusia and to right anterior MTG/inferior temporal gyrus (ITG) and hippocampus atrophy in rhythm amusia (Sihvonen et al., 2016; Sihvonen, Ripollés, et al., 2017). Little is known about the impact of acquired amusia on sensitivity to tonality, especially regarding its recovery over time.

The present study reported a case of tonal amusia combined with pitch amusia after stroke in the right temporal lobe but with intact rhythmic and low-level pitch processing ability. A series of assessments were conducted to evaluate the patient’s cognitive, musical and language abilities, and his recovery after a 3-month period. The study also explored the potential benefits for those with acquired amusia of combining melodies with corresponding visual cues to aid in their melodic processing.

Materials and methods

Patient information and relevant history

MM, a 62-year-old right-handed man, had a right middle cerebral artery (rMCA) infarction in December 2019 resulting left-sided sensory loss and weakness, involving the face, arm and leg. He was treated with mechanical embolectomy, to retrieve the thrombus from his proximal right middle cerebral artery. Approximately 2 months later, MM’s Cranial nerves and visual fields were clinically normal. Tone, power, coordination, sensation and fine motor control were normal in the limbs, even in the left hand. However, left-sided deep tendon reflexes were slightly brisker than those on the right, and the left plantar response was equivocal: both remnants of his right middle cerebral artery territory infarct.

As shown in Figure 1, brain Computed Tomography (CT) scans conducted 2 months later confirmed a large hypodense area in the right temporal lobe, a possible minor low-density change in the right thalamus, and a focal stenosis in the proximal intracranial left vertebral artery. Soon after the stroke, MM found he was suffering from symptoms of amusia, and later reported this to a neurologist. MM was an active recreational musician, with more than 10 years of academic training in music performance. In addition to childhood training in piano performance beginning when MM was 4-years old, MM engaged in 7 years of private lessons in violin performance.

Figure 1. Sagittal (panel a), axial (panel b) and coronal (panel c) CT (computed tomography) scans performed 2 months after the infarct. They demonstrate the right temporal & inferior frontal lobe infarct (dark region as directed by vertical arrows). The right superior temporal gyrus (STG), the right middle temporal gyrus (MTG) and the right insula were damaged by the stroke, but the right inferior temporal gyrus (ITG) was relatively spared. The infarct involves the right transverse temporal gyrus of heschl (×), which is the primary auditory cortex. The right striatum (S) is also shown, medial to the infarct. The left STG, left MTG, left ITG, left insular cortex (I), and left frontal lobe (F) are shown. The white region within the left STG is caused by the artifact from the left cochlear implant. A similar artifact (art) is also seen in the axial image (panel b).

MM had a profound sensorineural hearing loss in his left ear that occurred approximately 16 months prior to stroke. As shown in Figure 2, brainstem auditory evoked potentials conducted in June 2023 confirmed normal brainstem responses when the sound was presented in MM’s right ear whereas no discernible waveforms can be observed when the sound was presented in his left ear. MM daily used a cochlear implant (CI) to treat the left-ear tinnitus in early mornings and the CI was switched off after the treatment. MM typically removed the external processor associated with his CI for musical activities (citing poor musical perception through the device). The CI was not activated during the musical testing procedures reported in this paper. In fact, all auditory stimuli were presented exclusively to the normal-hearing right ear, so MM can be regarded as representing a single-sided deaf listener for these assessments.

Figure 2. MM’s auditory brainstem responses (ABR) performed in June 2023. The ABR in the right ear (panel a) shows normal responses and the latencies for wave I, wave III and wave V are 1.7 ms, 3.6 ms, and 5.3 ms whereas for the ABR in the left ear (panel b), no discernible waveforms can be observed.

Vignette from MM

Before my 2019 stroke, I was regularly engaged in musical performance, enjoying monthly rehearsals with a blues combo in which I played the bass guitar. My performance with the combo for a full year prior to the stroke was relatively effortless despite the sudden sensorineural hearing loss in my left ear that I experienced more than a year earlier. Then, after my 2019 stroke, my performance on bass guitar was significantly degraded at a rehearsal with those same musicians, who also found my performance noticeably poorer than usual. After becoming aware of this loss in my musical skill, I also noticed that I had great difficulty singing in tune when sitting at the piano. One particular event was particularly poignant in this regard. I found I was unable to sing the “Happy Birthday” song that year, in stark contrast to the sophisticated performance of which I was capable in the previous year (which included simultaneously playing the piano and singing the melody in harmony with my wife). After a great deal of practice, I succeeded in singing the entire song with the correct pitch on the next attempt for my grandson.

During the 3 months between the initial assessment and the follow-up assessment, MM engaged in substantial musical activities that included music composition, piano and guitar performance and multi-track recording. In addition, MM completed daily structured listening exercises that resemble two tasks more typical of experimental research in perceptual psychology. These two tasks were distinguished from more conventional musical activities most specifically by their structure as scientific tests of the listener’s ability to respond to brief musical stimuli that were presented in a randomized order under computer control. The first of these two tasks required MM to identify the musical interval formed by two simultaneously heard notes that could be heard to form a perfect fourth, a tritone, or a perfect fifth. The second task required the listener to identify which of four minor scales (i.e., minor modes) was exemplified by a melodic sequence of nine notes that were played beginning with three different tonic notes (A, D, and F). A comprehensive description of the experimental stimuli and methods is beyond the scope of the current paper.

Initial assessments

Neuropsychological assessment

Fortuitously, MM had a pre-stroke neuropsychological assessment in February 2018, following subjective memory complaints, which reported intact cognition in all areas, particularly in memory, which was found to be in the superior range. The tests of the neuropsychological assessment conducted 7 months post stroke, were selected based on MM’s specific cognitive concerns, and based on the battery of the prior neuropsychological assessment to allow comparison of results. Raw scores were converted to standardized scores in line with the administration and scoring instructions of each test. The results of the assessment were compared to the results of MM’s 2018 assessment. A significant change in scores over time was defined as a difference of more than one standard deviation (SD) from the 2018 baseline assessment.

All tests were administered by a Provisional Psychologist and a Senior Clinical Neuropsychologist. The test battery included Information and Orientation, and Mental Control from the Wechsler Memory Scale-III (WMS-III; Wechsler, 1997); Block Design, Similarities, Digit Span and Coding, from the Wechsler Adult Intelligence Scale-IV (WAIS-IV; Wechsler, 2008); Logical Memory I and II, and Symbol Span from the Wechsler Memory Scale-IV (WMS-IV, Wechsler, 2009); Verbal Fluency and Colour Word Interference from the Delis-Kaplan Executive Functioning System (D-KEFS; Delis et al., 2001); Trail Making Test A and B (TMT A&B; Reitan, 1958); Boston Naming Test (BNT; Kaplan et al., 1983); Simple Figure Copies; Clock Drawing; Line Orientation from the Repeatable Battery for the Assessment of Neuropsychological Status (RBANS; Randolph, 1998); Rey Complex Figure Test (RCFT; Rey, 1941); California Verbal Learning Test-Second Edition (CVLT-II; Delis et al., 2000); Wisconsin Card Sorting Test (WCST; Grant & Berg, 1948); and the Depression Anxiety Stress Scales − 21 items (DASS-21; Lovibond & Lovibond, 1995).

Musical assessment

Montreal Battery of Evaluation of Amusia

To confirm his self-reported amusia, MM’s musical deficits were assessed using the Montreal Battery of Evaluation of Amusia (MBEA; Peretz et al., 2003), which can be used in combination with other cognitive tests to establish a diagnosis of amusia (Vuvan et al., 2018). The MBEA comprises six subtests. The first three melodic subtests were developed to identify individuals with deficits in pitch processing. In each melodic subtest, listeners were presented with two melodies per trial, which were either identical or slightly different in terms of the pitch height of one note. The listeners were asked to judge whether the two melodies were the same or different. The Rhythm subtest was very similar to melodic subtests, but the difference between the two melodies, if any, was manipulated by modifying the duration of one note. The Meter subtest contained only one melody in each trial. The melody either had a March metric structure (duple meter) or a Waltz metric structure (triple meter). The listener was required to identify the meter of the melody. The last subtest was Memory where the listener needed to tell whether the melody was presented in the previous five subtests.

Pitch discrimination threshold

MM’s pitch discrimination threshold was measured by an adaptive “3-down-1-up” staircase paradigm (Levitt, 1971) which was used in a previous study (Sun et al., 2017). In each trial, MM was presented with three pure tones, and two of them were identical and the other one was either higher or lower than those two in pitch. MM was required to indicate which of the three tones was different from the other two.

Tonality test

To assess MM’s sensitivity to the tonality, we applied an out-of-key detection task used in our previous study (Sun et al., 2018), which was originally taken from Peretz et al. (2008). In each trial, MM was presented with a single melody, and judged whether the melody contained an out-of-key note.

Rhythm discrimination test

In order to determine whether MM’s low performance in the Rhythm subtest of MBEA (Peretz et al., 2003) was impacted by his pitch-related impairments, we tested him with Grahn and Brett (2009) rhythm discrimination test where the pitch of the tones was fixed within the rhythmic sequences. For each trial, three rhythmic sequences were presented. The first two sequences were identical, and the third sequence was either the same or different from the prior two in rhythm. MM was then asked whether the third sequence was the same or different from the first two. The rhythmic sequences can be grouped into the simple condition where the tones were arranged in a way that a regular meter can be perceived, and the complex condition where the tones were scrambled and no regular meter can be perceived (for more details, refer to Grahn & Brett, 2009; Sun et al., 2017).

Follow-up assessment

After the initial assessment session, informal interviews were conducted with MM. He noted specific difficulties when following compositions with multiple instruments, such as group-based music and ensembles. We believed this was likely due to difficulties with auditory stream segregation – the ability to perceive a sound source such as a piano into a coherent and perceptually meaningful unit. Hence, we included additional assessments to explore this. The follow-up assessment occurred after 3-months, comprised of the initial test battery and the following additional tests.

Additional musical assessment

Music in noise

The Music-In-Noise Task (MINT; Coffey et al., 2019) was used to test the patient’s auditory stream segregation abilities in a musical context. In the MINT task, each trial contains two melodies, which can be either identical or different from each other. One of the two melodies was presented with the multi-music noise, and the other one was presented in silence. The listener was required to identify whether these two melodies were the same or different regardless of the noise background. To explore influences of additional auditory and multimodal cues on task performance, the MINT was designed with five conditions: Baseline, Predication, Rhythm, Spatial and Visual. In the Baseline condition, the listener heard the melody with the noise background first followed by the comparison melody in silence. The other four conditions were the same as the Baseline condition except one manipulated point for each. In the Predication condition, the order of the melodies in each trial was reversed, and that is the melody in silence came first and the melody with noise was presented second. In the Rhythm condition, the melody was replaced by a rhythmic sequence with fixed pitch height. The Spatial condition introduced a cue pointing to left/right channel that the melody with noise background coming from. The Visual condition introduced a visual assistance to visualize the pitch height of each tone in the melody to aid the melodic discrimination.

Other auditory assessment

Spectral resolution

Given that MM had left-side hearing loss and wore a CI, we wanted to test if hearing acuity in the normal-hearing right ear was affected. In particular, we tested spectral resolution, which has positive correlations for music and speech perception for normal-hearing listeners and CI recipients (Choi et al., 2018). In order to measure the patient’s spectral resolution, the Spectral-temporally Modulated Ripple Test (SMRT; Aronoff & Landsberger, 2013) was used. SMRT is a 1-up-1-down adaptive forced choice task with three pure tones presented per trial. In each trial, two of the tones were identical and one was different from the other tones. The listener needed to identify which tone was different from the other two.

Speech in noise

To evaluate MM’s speech-in-noise perception, the Australian Sentence Test in Noise (AuSTIN; Dawson et al., 2013) was used. AuSTIN is an adaptive speech-in-noise test. A list of 20 sentences was randomly selected from the provided lists and spoken by a female speaker in the presence of time-locked four-talker babble noise. The initial signal-to-noise ratio (SNR) was 10 dB. If MM repeated more than 50% of the sentence correctly, the SNR would decrease, making the signal softer, and the task more difficult. If MM repeated less than 50% of the sentence correctly, the SNR would increase, making the signal louder, and the task easier. The first four sentences would adjust by 4 dB steps, and the remaining sentences by 2 dB steps. The speech reception threshold was calculated where the signal-to-noise ratio at which 50% of syllables were correctly perceived.

Procedure

All the assessments were conducted in a clinic room at the Royal North Shore Hospital, St Leonards, NSW Australia. Auditory-related tests were conducted separately from the cognitive tests. For the MBEA (Peretz et al., 2003), the pitch discrimination test (Sun et al., 2017), the tonality test (Peretz et al., 2008), and the rhythm discrimination test (Grahn & Brett, 2009), a headphone was used to deliver the stimuli at a comfortable sound level adjusted by MM. For the MINT (Coffey et al., 2019), SMRT (Aronoff & Landsberger, 2013), and AuSTIN (Dawson et al., 2013) tests, two loudspeakers were used considering MM’s left-side hearing loss and the fact that the stimuli of these tests can be different from the left and right channels. These stimuli were presented at 65 dB as measured with a sound-level meter from the patient’s listening position. For each test, a practice was provided for MM to get familiar with the task. Short breaks were available and provided as MM requested. The ethical approval was obtained from the Human Research Ethics Committee of Northern Sydney Local Health District. MM gave his consent for publication.

Results

Neuropsychological assessment

As shown in Table 1, MM’s neuropsychological assessment results were in line with or above normal limits for his age in all tests administered, despite performance discrepancy in some tests. Specifically, at baseline, MM performed in the 70^th percentile on Block Design (Wechsler, 2008) and in the 50^th percentile at his 2020 review, resulting a change of 0.55 SD. This would be considered as a normal variation in performance under test conditions, being under the one SD threshold and would thus not represent either an impairment or a clinically significant change in visuospatial skills. This is further supported by his performances on other tasks of visuospatial and visuoconstructional skills—Rey Complex Figure Test (Rey, 1941), Simple Copies, Clock Drawing and RBANS Line Orientation (Randolph, 1998), where his performance remained stable from his 2018 assessment and placed in the normal range or higher for his age. Overall, there is no evidence to suggest cognitive impairment in this domain, as defined by the DSM-V-TR (American Psychiatric Association, 2022)—a performance of below the 16^th percentile (minor neurocognitive impairment) or below the 2^nd percentile (major neurocognitive impairment).

Table 1. MM’s cognitive assessment results.

Domain	Pre stroke (February 2018)	After Stroke (July 2020)	Significance
Subtest	Percentile	Percentile	(z-score SD change)
Premorbid Estimate
Test of Premorbid Function	91	N/A	N/A
Orientation
Information/Orientation (WMS-III)	NAD	NAD	n.s
Attention & Working memory
Digit Span FWD (WAIS-IV)	75	75	n.s
Digit Span BWD (WAIS-IV)	91	63	−1*
Mental Control (WMS-III)	95	98	n.s
Information Processing Speed
Coding (WAIS-IV)	98	91	n.s
TMT-A	>90	90	n.s
Symbol Span (WMS-III)	N/A	91	N/A
Language
BNT	90	92	n.s
Visuospatial & Visuoconstructional
Block Design (WAIS-IV)	70	50	n.s
Copy (RCFT)	>16	>16	n.s
Simple Copies	N/A	NAD	N/A
Clock Drawing	N/A	NAD	N/A
Line orientation (RBANS)	N/A	88	N/A
Verbal Memory
Wordlist T1–5 (RAVLT)	84	99	N/A
Short delay free recall (CVLT-II)	98	99	N/A
Long delay free recall (CVLT-II)	98	99	N/A
Logical Memory I (WMS-IV)	84	50	−1*
Logical Memory II (WMS-IV)	91	50	−1.33*
Visual Memory
30 m delay (RCFT)	98	>99	n.s
Executive Function
Similarities (WAIS-IV)	50	50	n.s
Verbal Fluency (D-KEFS)
Letter	95	91	n.s
Category	84	50	−1*
Category Switching	63	91	1*
Switching Accuracy	75	91	n.s
Colour Word Interference (D-KEFS)
Colours	N/A	91	N/A
Word	N/A	91	N/A
Inhibition	N/A	91	N/A
Inhibition/Switching	N/A	95	N/A
TMT-B	75	91	n.s
WCST-64
Total Errors	N/A	42	N/A
Perseverative Errors	N/A	13	N/A
Categories completed	N/A	>16	N/A
WCST-128
Total Errors	1	N/A	N/A
Perseverative Errors	1	N/A	N/A
Categories Completed	2–5	N/A	N/A
Psychological Wellbeing – DASS-21
Depression	Normal	Mild	N/A
Anxiety	Moderate	Normal	N/A
Stress	Moderate	Moderate	N/A

BNT = Boston Naming Test (Kaplan et al., 1983).

CVLT-II = California Verbal Learning Test Second Edition (Delis et al., 2000).

DASS-21 = Depression Anxiety Stress Scales − 21 Items (Lovibond & Lovibond, 1995).

D-KEFS = Delis-Kaplan Executive Functioning System (Delis et al., 2001).

NAD = No Abnormalities Detected.

N/A = Not Applicable.

n.s = Not Significant.

RAVLT = Rey Auditory Verbal Learning Test (Rey, 1941).

RBANS = Repeatable Battery for the Assessment of Neuropsychological Status (Randolph, 1998).

RCFT = Rey Complex Figure Test (Rey, 1941).

SD = Standard Deviation.

TMT = Trail Making Test (Reitan, 1958).

WAIS-IV = Weschler Adult Intelligence Scale Fourth Edition (Wechsler, 2008).

WCST-64 = Wisconsin Card Sorting Test 64 Card Version (Grant & Berg, 1948).

WCST-128 = Wisconsin Card Sorting Test 128 Card Version (Grant & Berg, 1948).

WMS-III = Weschler Memory Scale Third Edition (Wechsler, 1997).

WMS-IV = Weschler Memory Scale Fourth Edition (Wechsler, 2009).

Percentile Classification.

>98 – Very Superior.

92–97 – Superior.

77–91 – High Average.

25–50 – Average.

9–23 – Low Average.

2–8 – Borderline Impaired.

<2 – Impaired.

Similarly, there were no impairments in naming or information processing, with performance not indicative of significant change across 2018 and 2020 assessments. MM performed in the 90^th percentile (2018—high average) and 92^nd percentile (2020—superior range) on the BNT (Kaplan et al., 1983), in the 98^th percentile (very superior) and 91^st percentile (high average) on WAIS-IV Coding (Wechsler, 2008), in the 91^st percentile (high average) on Symbol Span (WMS-IV, Wechsler, 2009), and 90^th percentile (high average) on TMT A test (Reitan, 1958).

There was a significant decline from the 2018 to 2020 assessment in only four tests: Digit Span Backwards (Wechsler, 2008), Logical Memory I and II (Wechsler, 2009) and Category Fluency (Delis et al., 2001). The changes in these tests were of between 1 to 1.33 SDs resulted in MM going from the high average to the average range and therefore considered a statistically significant but only mild decline from the baseline assessment, all of which still remained firmly in the normal range for someone of MM’s age. Clinically, these changes are not considered an impairment as per DSM-V-TR (American Psychiatric Association, 2022) when compared to the population mean. They also do not fit the pattern of cognitive impairment and profile of any neurodegenerative condition and are very likely to be sequelae from the rMCA stroke. Moreover, such minor changes in cognitive function cannot account for his severe musical deficits following the rMCA stroke.

Notably, these mild changes in working memory and information processing capabilities occurred when encoding and later recalling large amounts of contextualized information read as short stories. These, however, remained well within normal limits for an individual of his age and are not indicative of an impairment. Overall, these neuropsychological findings provide no evidence of any ongoing neurodegenerative disorder, and remain in the normal range so cannot account for his striking musical impairments.

Musical assessments

MBEA

As shown in Figure 3a, initially, MM’s average score across the three melodic subtests (Scale, Contour, and Interval) was 61%, which was below the cutoff score—72% obtained from the normal population for congenital amusia identification (Sun et al., 2017). MM’s Rhythm score was equal to the cutoff score (77%, Peretz et al., 2003). MM’s memory score (63%) was below the cutoff (73% Peretz et al., 2003). In contrast, MM scored 90% on the meter subtest high above the cutoff (67%, Peretz et al., 2003).

Figure 3. MM’ performance on the (panel a) MBEA (Peretz et al., 2003), (panel b) Grahn and Brett (2009) rhythm discrimination task, and (panel c) music-in-noise test (Coffey et al., 2019). For panel b, the control average scores were obtained from Sun et al. (2017). The arrows indicate the noticeable difference.

After 3 months, MM’s pitch impairments did not significantly improve. His melodic MBEA score increased slightly from 61% to 64% but still in the diagnostic range of amusia. MM’s Rhythm score was the same—77% at both time points. His Memory score remained in the amusic range and slightly reduced from 63% to 60%. On the Meter subtest, surprisingly, MM showed a noticeable decrease from 90% to 70% but still in the normal range.

Pitch discrimination threshold

MM’s initial pitch discrimination threshold (7.66 cents) was in the normal range, and remained in the same range (7.83 cents) after 3 months.

Tonality test

After 3 months, MM’s ability to detect out-of-key tones was significantly improved from 63% to 96%. See Figure 3a.

Rhythm discrimination test

D-prime scores (henceforth d’) were calculated based on the hit and false alarm rate obtained on each condition (Simple or Complex) of the rhythm discrimination test (Grahn & Brett, 2009). In the initial assessment, MM’s d’ score was 2.9 in the Simple condition, and 2.4 in the Complex condition, showing a bias to regular metric structure. After 3 months, this bias was gone—his score in the Simple condition was reduced to 2.2 which was the same with the score in the Complex condition (see Figure 3b). Compared to the control participants reported in our previous study (Sun et al., 2017), MM’s performance was above the average level except for his score in the Simple condition after 3 months, which was 0.4 SD below the average.

MINT

MM’s MINT average score was 61% which was equivalent to the 20^th percentile based on the norm reported by Coffey et al. (2019). As shown in Figure 3c, MM scored at the chance level (50%) in the Baseline condition. Removing the pitch variation of the melody (i.e., all tones in the melody have the same pitch height), as designed in the Rhythm condition, did not help him (50%). In the Spatial condition, a visual cue presented on the screen indicating the left/right channel where the target melody would come from slightly increased his accuracy to 55%. In the Prediction condition, where it presented the melody in silence prior to the melody with noise, MM did a bit better than the previous three conditions—65% but still far below the average level (i.e., 83%; Coffey et al., 2019). MM’s performance in the above four subtasks of MINT was below the average level by 1.4–2.8 SD based on the norm reported by Coffey et al. (2019). In contrast, as shown in Figure 3c, in the Visual condition where some boxes on the screen showed the pitch height and duration, he got 85% correct responses which was 0.25 SD higher than the average (Coffey et al., 2019).

Additional auditory assessments

MM’s spectral resolution was 6.2 ripples per octave in the SMRT test, which is within the expected range (Aronoff & Landsberger, 2013). MM’s speech reception threshold was −6.4 dB in the AuSTIN test, which means he was able to achieve 50% accuracy for sentences when the signal was 6.4 dB lower than the masking noise. This is better than the expected range for normal-hearing listeners (Choi et al., 2018).

Discussion

We conducted a series of auditory and cognitive assessments on a right-handed musician (MM) who developed amusia after a rMCA infarction. We found that the stroke leading to malfunction of the right temporal lobe mainly impacted MM’s musical ability, whereas language abilities remained within the normal ranges. This is consistent with previous findings that malfunction of the right temporal lobe can cause musical impairments (Hyde et al., 2007, 2011) but it may not influence language processing beyond the auditory analysis required (Gandour et al., 2004).

MM’s mild decline in working memory and information processing are in line with Särkämö et al (2009, 2010) findings that amusics with a left or right MCA stroke performed worse than non-amusics in working memory and some other cognitive areas. However, Rosemann et al. (2016) further investigation with amusics having small infarctions in MCA suggested that working memory deficits can be observed in these amusics, but these were not correlated to their musical impairments. Therefore, our findings support the view that amusia is not connected to other deficits in cognitive functions. MM’s mild changes in working memory and information processing may be mediated by lesions damaged areas responsible for different functions. As MM had left-side hearing loss which has been considered as a risk factor of cognitive decline, therefore, MM’s mild cognitive decline might also be associated with his hearing loss (Livingston et al., 2020).

MM’s pitch processing was impaired as reflected by the melodic MBEA (Peretz et al., 2003) and there was no improvement over the 3 months. MM showed difficulty learning and remembering novel melodies. As MM performed in the normal range on the general attention, working memory, information processing, and verbal learning and memory tests as measured in the cognitive assessments, we suspect that the melodic memory problem may be the consequence of the deficit in pitch processing. MM’s poor performance in the Rhythm subtest of MBEA may largely reflect his pitch deficit, whereby pitch variation in melodies made the rhythmic task harder (Sun et al., 2017). Previously acquired amusia studies have found that lesions in the right temporal area were associated with poor performance on the Scale and Rhythm subtests of MBEA (Sihvonen et al., 2016; Sihvonen, Ripollés, et al., 2017). Such associations may also arise from the interaction between pitch and rhythm required in the Rhythm subtest. Indeed, MM’s rhythm discrimination test (Grahn & Brett, 2009) result revealed his rhythmic processing was normal when the pitch height of the melodies was held constant. MM’s tonality processing was impaired but it recovered to normal after 3 months. While MM initially had good metric perception, it declined after 3 months. We also found evidence to support MM’s anecdotal suggestions around the difficulty of processing melodies in the presence of a musical noise background. To the best of the authors’ knowledge, we present the first finding that visual aids can assist a person with amusia in their perception of music-in-noise.

All of these musical impairments cannot be explained by low-level auditory deficits, as MM had normal pitch discrimination thresholds and spectrotemproal modulation thresholds. Although MM has a single-sided deafness typically associated with music sounding less natural, pleasant, and distinct, these are generally the result of the loss of stereo sound (Meehan et al., 2017). In our case, during testing, MM removed his CI and did not experience any tinnitus. All auditory tasks were presented unilaterally to the normal-hearing ear. MM had a history of musicianship after receiving the CI and only noted music perception difficulties post-stroke. This reiterates the onset of amusia being associated with the stroke, rather than deficits associated with perception through a CI.

MM showed difficulty processing tonality during his initial assessments, in that he had difficulty identifying out-of-key melodies. He also reported that it was difficult to identify whether a novel melody was in tune or out of tune. After 3months of recovery, however, MM could identify the out-of-key melodies with high accuracy. To the author’s knowledge, this is the first report of recovery from tonal amusia. During the 3 months, MM engaged in conventional musical activities and listening exercises focused on tonality and perception of musical intervals, which likely accelerated MM’s recovery in tonal processing. In contrast, no improvement was seen in his melodic processing as reflected by the melodic MBEA subtests (Peretz et al., 2003). This finding is consistent with previous evidence that acquired amusia with a lesion in the right temporal lobe is associated with poor recovery of melodic processing abilities (Sihvonen et al., 2016; Sihvonen, Ripollés, et al., 2017). This finding further suggests the dissociation between tonal processing measured by the out-of-key detection task (Peretz et al., 2008) and the melodic processing assessed by the melodic subtests of the MBEA (Peretz et al., 2003).

Initially, MM’s amusia did not impact his metric perception. He performed well in the Meter subtest of the MBEA (Peretz et al., 2003). This can also be reflected by MM’s bias to the Simple condition rather than the Complex condition in Grahn and Brett (2009) rhythm discrimination test, which means he was able to perceive the regular metric structure of the tone sequences and utilize it to help him remember and discriminate the rhythmic differences. This finding indicates a dissociation between the rhythm perception and metric perception, which has been observed in acquired amusia (Liégeois-Chauvel et al., 1998). Theoretically, rhythm is a pattern of discrete durations depending on the underlying perceptual mechanism of grouping (Clarke, 1999), whereas meter is a pattern of beats at isochronously spaced intervals, which allows for rhythmic expectations in music (Honing, 2012, 2013; Vuust & Witek, 2014). Processing rhythm and meter may involve different brain areas. Neuroimaging research in healthy population found that compared to rhythm with irregular metric structures, rhythm with regular metric structures elicited increased activity in the basal ganglia and supplementary motor areas (SMA)/pre-SMA (Grahn & Brett, 2007).

Unexpectedly, MM’s metric perception considerably declined after 3 months. His Meter score on the MBEA (Peretz et al., 2003) significantly reduced from a high above average level to a slightly above the cutoff score level, and he was not able to benefit from the metric structure when engaged in Grahn and Brett (2009) rhythm discrimination task, although he still can discriminate between rhythmic sequences at a normal level. Pitch salience theory may help to explain this reduced sensitivity to meter over time (Prince et al., 2009). In the initial assessment, melodic and tonal impairments were evident, reducing the perceptual salience of pitch information in music. Reduced pitch salience, in turn, may have allowed metric structure to capture attention. As MM’s sensitivity to tonal structure improved over time, pitch salience was reestablished, reducing the relative salience of metric structure and revealing the underlying deficit.

In the present case study, we tested the music-in-noise skill of the individual with acquired amusia. To the best of our knowledge, this is the first time a music-in-noise test has been used to test for amusia. Our amusic patient MM responded at a chance level to most conditions on this task. This suggests impaired performance in this task may be related to musical disorders such as amusia. However, MM’s music-in-noise skill was significantly improved when visual cues (along with the music) showed the pitch height and duration of each note in the music. Visual cues assist in segregating melodies from distracting maskers (Marozeau et al., 2010). However, this effect may be particularly pronounced in MM due to his hearing loss and use of a cochlear implant. It is well established that many individuals with hearing loss utilize visual cues to compensate for diminished auditory acuity. Hence, they display superior lip-reading skills relative to normal-hearing peers (Pimperton et al., 2017).

Conclusion

This study reported a case of comorbid tonal amusia and pitch amusia caused by a rMCA infarction. The patient’s sensitivity to tonality recovered after 3 months with musical engagement. To the author’s knowledge, this is the first report of recovery from tonal amusia. Further research is needed to investigate tonal amusia with a larger number of participants, and its rehabilitation and possible training. The present study supports the existence of a variety of musical disorders (Peretz et al., 2003) and it supports the modality model of music processing whereby musical processing may involve in a complex brain network, with different brain areas for different components of musical processing such as pitch discrimination, tonal processing, and rhythmic processing (Peretz & Coltheart, 2003). Therefore, these musical abilities can be selectively affected in people with brain injury. Our findings also provide the first evidence of visual aids enhancing perceptual abilities for an amusic. Further research is needed to explore the possibility of visual enhancement of musical cues as a strategy to improve the perceptual deficits associated with amusia.

Disclosure statement

No potential conflict of interest was reported by the author(s).

Data availability statement

Raw data can be shared under reasonable request by e-mail to the correspondent author.

Additional information

Funding

This research was supported by the Australian Research Council via a discovery project grant awarded to WFT [DP210101247].