Frontal Verbal Amnesia
Phonological Amnesia
H. Branch Coslett, MD; Dawn Bowers, PhD; Mieke Verfaellie, PhD; Kenneth M. Heilman, MD
\s=b\ Although aphasic patients with frontal
lobe damage may demonstrate impaired retention of verbal material, significant anterograde memory disturbances have not, to our
knowledge, been reported with a minor Broca's aphasia. We describe a patient with minor Broca's aphasia who exhibited an unusual and profound anterograde memory
disturbance, especially for phonologically
specified stimuli. We suggest that this disturbance is attributable to an impairment in the
volitional, controlled search of stored phonological information.
(Arch Neurol. 1991 ;48:949-955)

number of investigators have dem"^ onstrated that anterior
aphasia
may be associated with impaired reten¬
tion of verbal material.1"3 Cermak and
Moreines,4 for example, reported that
patients with nonfluent aphasia were
impaired on a phonological memory task
involving the detection of rhymes; they
proposed that anterior aphasia induced
a defect of "phonemic recirculation."
More recently, Cermak and coworkers0
have further suggested that the memo¬
ry deficits of anterior aphasies may be
related to an inability to compare phono¬
logical items already in memory and dif¬
ficulty in "reconstructing" items based
on phonological features.
Broca's aphasia is typically associated
with lesions involving Broca's area as
well as portions of the anterior perisyl¬
vian region, including the sensory mo¬
tor cortex, insula, and anterior parietal
operculum." Smaller lesions largely or
entirely restricted to Broca's area, how¬
ever, have also been shown to be associ¬
ated with a language disturbance; this
syndrome, designated minor Broca's
aphasia by Foix,7 is characterized by a
combination of difficulties, often includ¬
ing word-finding disturbance, effortful
articulation, and decreased verbal flu¬
ency. Although sometimes initially se¬
vere, these symptoms are typically
transient, and persistent, overt lan¬
guage disturbances are uncommon.

Accepted for publication January 31,1991.

From the Department of Neurology, Temple
University School of Medicine, Philadelphia, Pa
(Dr Coslett), and the University of Florida College
of Medicine and the Veterans Affairs Medical Center, Gainesville, Fla (Drs Bowers, Verfaellie, and

Heilman).

Reprints not available.

Although nonaphasic patients with
left frontal lobe lesions may have lexical
retrieval impairments, as evidenced by
their poor performance on controlled
word association tasks,8,9 to our knowl¬
edge significant anterograde memory
disturbances have not been reported
with a minor Broca's aphasia. We de¬
scribe a patient with minor Broca's
aphasia subsequent to an ischemie in¬
farction involving Broca's area and
small portions ofthe inferior and middle
frontal gyri who exhibited an unusual
and profound anterograde memory dis¬
turbance; we suggest that this distur¬
bance is attributable to an impairment
in the volitional, controlled search of
stored phonological information.
REPORT OF A CASE

The patient was a right-handed 59-yearold man who was admitted for evaluation of a
memory impairment. He had been in excel¬
lent health until 3 years prior to admission
when, 1 week after a transient episode of lefteye blindness, he noted the sudden onset of
difficulty in speaking and weakness of the
right side of his body. Neurologic evaluation
at that time revealed a mild right hemipare¬
sis, a nonfluent aphasia with mildly impaired
auditory comprehension, and a Gerstmann's
syndrome. The auditory comprehension im¬
pairment and Gerstmann's syndrome re¬
solved within several days. Unfortunately,
no additional information about the patient's
initial language impairment is available.
Although family members noted no im¬
pairment in his speech after the stroke, the
patient considered himself to be appreciably
less articulate because he had difficulty in
finding the appropriate word and in "getting
words out."
In the estimation of the patient and his
family, however, his subtle speech impair¬
ment paled in significance relative to the se¬
vere and disabling memory deficit noted sub¬
sequent to his stroke. Although he believed
his memory for faces, pictures, and places to
be relatively unimpaired, he stated that his
memory for names and verbal material,
whether written or spoken, was quite poor.
As a consequence of his memory deficit, he
was unable to return to his job as a loan
officer for a finance company; he worked
briefly as a prison officer before his memory
disturbance forced him to retire.
Neurologic examination 3 years after the
stroke revealed a mild spastic weakness and
clumsiness of the right hand. Evaluation of
the patient's higher nervous system func¬
tions revealed that his spontaneous speech,
although marked by occasional word-finding
pauses, was fluent, with normal prosody and

Downloaded From: http://archneur.jamanetwork.com/ by a Michigan State University User on 06/11/2015

articulation. His discourse contained a nor¬
mal range of syntactic structures. Naming,
although occasionally hesitant, was normal.
Repetition of words, phrases, and sentences
was excellent; sentences of up to 12 words, as
well as sentences with unusual syntactic
structures, such as center-embedded

clauses, were repeated perfectly. Auditory
comprehension was unimpaired. The patient
read aloud and for comprehension quite well.
No effect of orthographic regularity, part of
speech, or imagery was noted in his reading.
His writing was normal. Calculations, rightleft discrimination, and finger naming were
normal. No buccofacial or limb apraxia was
noted. There was no neglect. He was able to
draw a cube, as well as maps of the United
States.

A computed tomographic scan performed
3 years after the patient's stroke revealed an
old area of infarction involving Broca's areas,
as well as more rostral portions ofthe inferior
frontal gyrus. Additionally, small portions of
the precentrai and middle frontal gyri were
also infarcted (Figs 1 and 2).

NEUROPSYCHOLOGICAL EVALUATION
Intellectual

On the Weschler Adult Intelligence

Scale, the patient obtained a verbal in¬
telligence quotient (IQ) of 93, a perfor¬
mance IQ of 88, and a full-scale iQ of 91.
His performance IQ was thought to
have been adversely influenced by his

right-hand clumsiness. The Peabody
Picture Vocabulary Test10 yielded a ver¬
bal IQ estimate of 103.
Language

A number of formal tests were admin¬
istered for more careful assessment of
the patient's language capacities. His
performance on the Western Aphasia
Battery11 and selected portions of the
Boston Diagnostic Aphasia Examina¬
tion,12 including the reading, repetition,
and writing subtests, was essentially
perfect. For further assessment of the
patient's auditory comprehension, the
Token Test13was administered; the pa¬
tient made one mistake on subtest VI,
but otherwise his performance was per¬
fect. He performed normally on the Bos¬
ton Naming Test"; the Florida Syntax
Comprehension Test,15 in which correct
performance is critically dependent on
comprehesion of the distinction be¬
tween indirect and direct objects; and
the Battery of Adult Reading Func¬
tion,16 a test that assesses reading com¬
prehension and the reading aloud of reg-

Table 1.—Wechsler Memory Scale

(Form I)
Information
Orientation
Mental control
Memory passages

6

Digits total

8
5

5

6

6, 4 delayed

Visual reproduction
Associate learning

Memory quotient

1.5
71

mance was in the normal range on both

tasks.

Thus, the patient performed normal¬
ly on a wide range of language tasks and
exhibited a normal verbal IQ. In spite of
his at least relatively normal language
Fig 1.—A low ventricular computed tomo¬
graphic scan image demonstrating an infarct
involving the inferior frontal gyrus.

Fig 2.—A high ventricular computed tomo¬
graphic scan image demonstrating an infarct
involving the inferior frontal gyrus and small
portions of the middle frontal gyrus.

ular and irregular nouns, functors, and
nonwords. The Story Completion Test"
was also administered; with a single ex¬
ception, the patient generated a re¬
sponse ofthe desired syntactic form.

Scale was six initially (average normal,
10) and four after a 20-minute delay (av¬
erage normal, eight). His paired-asso¬
ciate learning was similarly impaired on
word-list learning tasks, both with audi¬
tory and visual presentation. On the
Rey Auditory Verbal Learning Test,22
for example, he recalled no more than
six of the 15 words across any of five
learning trials, and his delayed 20-min¬
ute recall was three. Similar results
were obtained with visual presentation.
His memory performance fell more than
2 SDs below the mean of age-matched
controls.
The patient's performance on the Rey
Auditory Verbal Learning Test also
provided an opportunity for assessment
of the effects of recency and primacy in
supraspan list learning. On the first five
trials on both the auditory and visual
versions of the Rey Auditory Verbal
Learning Test, the patient correctly
produced 90% of the 20 items presented

Frontal Lobe Tasks

Several cognitive tasks sensitive to
frontal lobe functioning were adminis¬
tered. The patient performed poorly on
the Wisconsin Card Sorting Test18; he
failed to detect the correct sorting prin¬
ciple after 35 trials. Approximately 20%
of his errors were random guesses and
50% were perseverative errors. The pa¬
tient was also impaired on the colornaming and interference subtests ofthe
Stroop Interference Task.19 He re¬
quired 88 seconds to name color patches
(1.5 SDs below the mean) and 288 sec¬
onds to complete the interference task
(2 SDs below the mean). The patient
demonstrated no perseveration in Lur¬
ia's "mn" writing task but performed
quite poorly on the "go-no go" and con¬

trasting motor programs task.2"
Attention and Memory

Digit span was assessed on a number
of occasions. The patient correctly re¬
peated 100% of orally presented fivedigit lists and 60% of six-digit strings.
Seventy-one percent of the errors were
sequencing errors, and 29% of the er¬
rors were due to the omission of a single
item. With visual presentation of singledigit written numbers at the rate of ap¬
proximately one number per second,
the patient's performance was quite
similar. Administration of the Wechsler
Memory Scale21 revealed the patient to
have a memory quotient of 71 (Table 1).
His story recall of the logical memory
stories from the Wechsler Memory

in the last two positions. Additionally,
the effects of recency and primacy were
evaluated by calculation of the percent¬
age of words reported as a function of
position in the list; in the first five trials
on the auditory and visual versions of
the test, the patient correctly reported
20% of the words from the first third,
16% ofthe words from the middle third,
and 72% ofthe words from the last third
ofthe list. Thus, the patient exhibited a
clear effect of recency but no clear effect
of primacy.
In contrast, the patient performed
much better on nonverbal recent memo¬
ry tasks such as the Milner Facial Rec¬
ognition Test23 and Kimura's Recurring
Figures Test.24 He correctly identified
nine ofthe 12 target faces and obtained a
score of 44 on the latter task; his perfor-

Downloaded From: http://archneur.jamanetwork.com/ by a Michigan State University User on 06/11/2015

skills, however, he has a marked verbal
memory deficit with sparing of nonver¬
bal memory. The following investiga¬
tions were performed to further charac¬
terize the patient's memory im¬

pairment.

Rate of

Forgetting.—As perfor¬
assessing the rate of for¬

mance on tests

getting has been demonstrated to dis¬
tinguish between amnestic syndromes
caused by bitemporal vs diencephalic
damage,20"27 the patient's rate of forget¬
ting was assessed in a manner similar to
that described by Huppert and Piercy.2"
One hundred target stimuli (pictures)
were individually projected on a screen
for a period of 8 seconds each. At inter¬
vals of 10 minutes, 20 minutes, 90 min¬
utes, 24 hours, and 1 week after this
presentation, different random subsets
of 20 pictures from the original sample
and 20 new pictures were projected, and
the patient was asked to indicate wheth¬
er each picture was from the original list
of stimuli. Five age- and educationmatched controls were tested in a simi¬
lar fashion, except that stimuli were
presented for 1 second each. Although
the difference in exposure times ren¬
ders a comparison with control subjects
difficult, it is apparent from examina¬
tion of Fig 3 that the patient's perfor¬
mance was quite similar to that of five
age- and education-matched controls,
falling within 1 SD of their performance
at every test interval.
The patient's rate of forgetting for
verbal stimuli was also assessed in the
manner described above. Stimuli for
this experiment included 100 slides de¬
picting a single typed word. The target
words were presented for 8 seconds
each, and the patient's recognition was
assessed in the manner described
above. The patient correctly sorted only
57% of the stimuli at both 10 and 20
minutes; as his performance was not
above the chance level ( 2 1.63; P>. 1),
this investigation was discontinued af¬
ter the second session. Errors included
=

.

an almost equal number of false-posi¬
tives and false-negatives. With a 1-second stimulus exposure, controls cor¬
rectly identified 81% and 75% of stimuli
at 10 and 20 minutes, respectively, and
remained well above chance (67% cor¬
rect) after 1 week (Fig 4).
Thus, the patient's rate of forgetting
was essentially normal for "nonverbal"
material; his rate of forgetting for ver¬
bal material could not be assessed, as we
could not demonstrate evidence of ade¬
quate initial performance.
Effect of Interference.—A Peterson
and Peterson28 task adapted from But¬
ters and Grady29 was given to assess the
effects of interference on the patient's
verbal memory performance. On each
trial, three semantically and phonologically unrelated monosyllabic or bisyllabic words were auditorally presented to
the patient at the rate of approximately
one word per second. Five trials were
presented in each of four conditions:
(1) immediate recall, in which there was
no delay between stimulus presentation
and production; (2) a 20-second unfilled
delay between stimulus presentation
and production, during which the pa¬
tient was permitted to rehearse; (3) a
20-second filled delay, during which the
patient was asked to count backward by
twos before recalling the stimuli; and
(4) a 4-second unfilled interval, during
which the patient was encouraged to
rehearse, followed by a 16-second peri¬
od, during which the patient was asked
to count backward by twos. In the im¬
mediate recall and 20-second unfilled
delay conditions, the patient correctly
produced all 15 ofthe target words; with
a 20-second filled delay, the patient pro¬
duced only two (13%) of 15 words; with a
4-second predistractor delay, followed
by a 16-second filled delay, he produced
nine (60%) of 15 words (Fig 5). On the
latter two conditions, errors included
16 omissions and two intrusions. No se¬
mantic errors were noted. The patient's
performance was quite similar with vi¬
sual presentation of the stimuli. He
made no errors in the immediate recall
and unfilled delay conditions, and he
correctly produced six (40%) of 15 stim¬
uli with a 20-second filled delay and pro¬
duced eight (53%) of 15 stimuli with a
4-second predistractor period and a
16-second delay and distraction. Errors
included three intrusions and 13 omis¬
sions.
These data suggest that the patient is
quite sensitive to the effects of interfer¬
ence or, alternatively expressed, de¬
pendent on continued rehearsal. In con¬
junction with the demonstration that
the patient exhibited a relatively nor¬
mal recency effect, these data suggest
that rehearsal within the articulatory

10 min

20 min

90 min

24 h

1 wk

Interval to Test

Fig 3. —Performance of the patient and controls, expressed as percentage correct, on a test
assessing the rate of forgetting of pictures.

10 min

20 min

90 min

24 h

1 wk

Interval to Test

Fig 4.—Similar to Fig 3 except that stimuli were words rather than pictures. NA indicates not
applicable.

Fig 5. —Percentage correct on a Peterson-Peterson memory task with auditory stimuli.
Patient

No Delay

20-s Silent Delay
Conditions

Downloaded From: http://archneur.jamanetwork.com/ by a Michigan State University User on 06/11/2015

20-s Delay/
Distraction

4-s Predistractor

Delay

100
80

8060

-Q- Patient, No Shift

40'

- - Patient, Shift
*- Controls, No Shift
-0- Controls, Shift

20'

-o- Patient

-*

Case

Rhyme

Controls

Semantic

Condition

Fig 7. —Performance, expressed as percent¬

age correct, of the patient and controls on a
level of processing task with written stimuli.
->—

-

6

Trials

Fig 6.—The performance of the patient and controls, expressed as percentage correct, on a
release from proactive inhibition task.

loop is not impaired.

Release from Proactive Inhibi¬
tion.—In an attempt to assess the pos¬
sible role of impaired semantic/concep¬
tual encoding in the production of the
patient's memory deficit, a releasefrom-proactive-inhibition task similar
to that of Cermak et al3 was adminis¬
tered. This task was derived from Wickens'30 modification of the Peterson and
Peterson28 distraction technique, in
which subjects were given several con¬
secutive trials of word stimuli, all of
which belonged to the same semantic
category (eg, fruits). This was followed
by a trial in which the word stimuli were
exemplars of a different category (eg,
vegetables). As previously document¬
ed, memory performance in normal sub¬
jects deteriorates over trials, because of
the buildup of proactive interference.

Wickens30 demonstrated, however,

a

dramatic improvement in memory per¬
formance on trials when the stimuli
came from a different semantic catego¬
ry. This improvement presumably
stemmed from a release of proactive in¬
terference that occurred because sub¬
jects had noted the change or shift in the
semantic category of the stimulus
words. The underlying assumption is
that stimuli must be semantically en¬
coded for the release effect to occur.
To assess release from proactive in¬
terference, the patient ana controls
were tested on a Peterson and Peter¬
son28 distraction task involving eight
blocks of five trials. Stimuli for each
trial consisted of three semantically re¬
lated concrete nouns (eg, pear, apple,
and grape) read aloud to the patient.
After a 20-second delay filled with
counting backward by twos, the patient
was asked to produce the three target
words. There were two conditions. In
the no-shift condition, all five trials of

the block consisted of word stimuli from
the same semantic category (eg, fruit).
In the category shift condition, word
stimuli on trial 5 were from a different
semantic category (eg, vegetables) than
the preceding four trials of the block.
Category shift and no-shift conditions
were randomized, with 5-minute rest
intervals between each block. The per¬
formance ofthe patient and normal con¬
trols across trials 1 through 5 of the
categoryshift and no-shift conditions is
shown in Fig 6. For both the patient and
the controls, there was a progressive
decline in performance across trials
1 through 4 of both conditions, with the

patient exhibiting a greater decrement,
which presumably reflects a greater
susceptibility to interference. The pa¬
tient, like amnesic patients with Korsakoff's psychosis,3 exhibited no "release"
effect. For controls, in contrast, the
change in semantic class on trial 5 ofthe
category shift condition was associated
with a significant increase in recall as
compared with the no-shift condition

(X2 4.91;ri/=1;P<.05).
=

The interpretation of this finding,
however, is unclear. Some investiga¬
tors3 have argued that the finding in¬
dicates that the patients were not en¬
coding in semantic categories. Alterna¬
tively, on the basis of a comparison of
the performance of patients with frontal
or temporal lobectomy on a releasefrom-proactive-inhibition task, Mosco¬
vitch31 has argued that the phenomenon
is more appropriately associated with
frontal impairment than with memory
loss.32 Because our patient was impaired
on tests of frontal lobe function as well
as memory, the explanation for his fail¬
ure to manifest the release phenomenon
is not clear.
Memory for Phonology.—Levels of
Processing.—Perhaps the most strik-

Downloaded From: http://archneur.jamanetwork.com/ by a Michigan State University User on 06/11/2015

ing aspect of the patient's verbal memo¬
ry deficit was a marked discrepancy be¬
tween performance on tasks requiring
registration and/or retrieval of phonologically, compared with semantically,
specified items. This phenomenon was
first noted on a task employing the "lev¬
els of processing" paradigm.33 This ex¬
periment was motivated by a conceptu¬
al framework suggesting that the
adequacy of memory registration is a
function of the depth of analysis or the
degree to which the stimulus is semanti¬
cally processed.
Stimuli for this task included 36 tar¬

get words printed on a 4 x 6-in card; half
of the words were printed in uppercase
and half in lowercase letters. Stimuli
were shown to the patient for approxi¬
mately 5 seconds each; immediately af¬
ter the presentation of each stimulus,
the patient was asked one of three types
of orienting questions: (1) one third of
the stimuli were accompanied by ortho¬
graphic orienting questions (eg, "Is this
word written in uppercase letters?");
(2) one third were accompanied by pho¬
nological orienting questions (eg, "Does
this word rhyme with fear?"); and
(3) one third were accompanied by se¬
mantic orienting questions (eg, "Is this
a type of fruit?"). Ten minutes later, a
recognition memory task was adminis¬
tered in which thirty-six 4 6-in cards,
each containing one of the stimuli and
three phonologically and semantically
unrelated foils, were presented. The pa¬
tient was asked to indicate which of the
four words had been previously

presented.
As depicted in Fig 7, the patient cor¬

rectly identified 87% ofthe words intro¬
duced with a semantic orienting ques¬
tion, 87% of the words introduced with
an orthographic orienting question, but
only 25% ofthe words introduced with a
phonological orienting question. In con¬
trast, five age- and education-matched

controls showed the expected pattern,
correctly identifying 91% of the seman¬
tically, 75% of the phonologically, and
67% ofthe orthographically cued words.

·

Table 2.—Fluency Tasks
Patient

Controls

Letters
f

2.4

a

2.0

s

2.6
12.3
2.0
16.0

Semantic categories
Rhyming words
Nonverbal (figures)

14.5
8.4

performed in which he was asked
to generate words beginning with a and
s. Data were collected for each of the
were

Delay
Trials

Fig 8. —Number of rhyming and semantically related paired associates recalled by the patient with
auditory and written stimuli.

The patient's selective memory deficit
for phonologically cued stimuli could not
be attributed to an inability to make
rhyme judgments, in that he was 100%
accurate when answering rhyme orient¬

ing questions.
An auditory version of the levels of
processing task was also performed. On
this task, the orthographic orienting
question was eliminated, and the pa¬

tient was asked to indicate if the words
were spoken in a male or female voice.
The patient correctly identified 93% of
the words presented with a semantic
orienting question, 75% of the words
introduced with a male/female voice ori¬
enting question, but only 25% of the
words introduced with a phonological
orienting question. Controls correctly
identified 91%, 75%, and 67% of stimuli
introduced with semantic, phonological,
and voice orienting questions.
Rhyming Paired-Associate Tasks.—
The patient's apparent deficit in memo¬
ry for phonologically encoded and/or re¬
trieved information was also assessed
using two list-learning tasks adapted
from Warrington and Weiskranz.34
Stimuli for one task included 12 pairs of
semantically related words (eg, "ask"
and "inquire"); stimuli for the second
task included 12 pairs of rhyming words
(eg, "chair" and "bear"). The word pairs
from each list were read to the patient at
the rate of one word per second. After
presentation of the paired-associate
list, the patient was given the first word
ofthe pair and asked to produce its asso¬
ciate. This procedure was repeated for a
total of four trials for each of the two
lists. Delayed recall of the word pairs
was tested after a 30-minute delay. The
semantically and phonologically related

word pairs were presented in different
sessions. Both visual and auditory ver¬
sions of these lists were given.
The patient's performance with writ¬
ten and auditory stimuli presentation is
presented in Fig 8. He performed better
with semantically related stimuli on
each trial with both written and audi¬
tory stimuli. With written stimuli, this
difference reached significance on trials
2, 3, and 4 ( 2 6.3, 8.7, 12.0, respec¬
tively; all P<.01). With auditory stimu¬
li, he performed significantly better
with semantically related pairs on trials
3 and 4 (x2 10.7and 8.2, respectively;
both P<.05). Summing across the five
trials, the patient performed signifi¬
cantly better with semantic, as com¬
pared with rhyming, paired associates
with both written ( 2 30.0; P<. 01) and
auditory ( 2 17.7; P<.01) stimuli.
Age- and education-matched controls
demonstrated no significant difference
between rhyming and semantically as¬
sociated words on individual trials but
did demonstrate a slight but consistent
superiority with semantic paired asso¬
ciates which, summing over all five tri¬
=

=

=

=

als, was significant ( 2 5.6; P<.05).
Fluency Tasks.—The effect of pho¬
nological factors on retrieval of informa¬
tion from remote or long-term memory
was assessed by a number of variants of
the Thurstone Word Fluency Test3,5 (Ta¬
ble 2). In one investigation, the patient's
ability to generate words beginning
with a designated letter was compared
with his ability to generate words de¬
=

fined on the basis of semantic attri¬
butes. In the former task, the patient
was instructed to generate in 1 minute
as many words as possible that begin
with the letter/; similar 1-minute trials

Downloaded From: http://archneur.jamanetwork.com/ by a Michigan State University User on 06/11/2015

three letters during five separate ses¬
sions. The patient was also instructed to
generate in 1 minute as many exemplars
as possible of a specified semantic cate¬
gory. Eight semantic categories were
tested on each of two occasions.
As indicated in Table 2, the patient's
average score for the letters f a, and s
was 2.4, 2.0, and 2.6 words per minute,
respectively; thus, his average across
all three letters was 2.3 words per min¬
ute. Performance with semantically de¬
fined categories was substantially bet¬
ter, as he generated an average of 12.3
items per minute, with a range of 6 to 18.
On another occasion, a word was pro¬
vided, and the patient was asked to gen¬
erate as many rhyming words as possi¬
ble in 1 minute. A total of five words
were presented. He produced an aver¬
age of 2.0 rhyming words per minute; he
did not produce nonrhyming words.
Age- and education-matched controls,
in contrast, generated an average of
8.4 words per minute.
In an effort to demonstrate that the
diminished fluency reflected a specifi¬
cally verbal impairment, a nonverbal
analog of the Thurstone Word Fluency
Test was administered in which the pa¬
tient was asked to create as many differ¬
ent drawings in 4 minutes as possible
using only four lines.36 On two trials he
created an average of 16 drawings, a
performance that is within the normal
range.
COMMENT

We have presented data derived from
59-year-old man with a minor Broca's
aphasia secondary to a documented left
a

dorsolateral frontal lobe infarct who
demonstrated a severe memory impair¬
ment of a type not previously described,
to our knowledge. Evaluation ofthe pa¬
tient's memory revealed that he was
sensitive to distraction, did not demon¬
strate the normal release from proac¬
tive inhibition, showed an effect of re¬
cency but not primacy, and, perhaps
most interestingly, exhibited a marked
memory impairment for the encoding
and/or retrieval of phonologically, com-

with semantically, specified
items. Several issues concerning this
patient's performance warrant atten¬
tion.
The first issue concerns the relation¬
ship between the patient's language ca¬
pacities and his memory impairment.
Abundant evidence suggests that both
fluent and nonfluent aphasias may be
associated with a verbal memory im¬
pairment. Might the patient's memory
deficit be secondary to a subtle language
deficit?
A number of specific mechanisms
have been proposed to account for the
impairments of short-term memory en¬
countered in aphasie patients. The im¬
paired verbal short-term memory en¬
countered in a probable Wernicke
aphasie, for example, has been attribut¬
ed to a failure to generate an adequate
phonological representation37; the im¬
pairment seen in certain conduction
aphasies has been attributed to a patho¬
logic reduction in the capacity of the
short-term store38,39; finally, in certain
Broca's aphasies, the impairment of
short-term memory has been attributed
to failure of repetition, with a resultant
failure of rehearsal in the putative artic¬

pared

ulatory loop.18,40

None of these proposed explanations,
however, can account for the data de¬
rived from our patient. Our patient ex¬
hibited no demonstrable phonological
impairment on tests of auditory com¬

prehension, phoneme discrimination,
and rhyme judgment. Additionally, our
patient had no impairment of repetition
and, as indicated by his perfect perfor¬

mance on the Peterson and Peterson28
task with an unfilled delay, was able to
retain phonologically encoded data, pre¬
sumably by means of rehearsal within
the articulatory loop.
Because the patient's infarct involves
Broca's area, a brain region thought to
be important in the control of motor
aspects of speech production, one might
suggest that the memory deficit reflects
an impairment at the level of motor pro¬
gramming for speech. There are several
lines of evidence that make this possibil¬
ity seem unlikely. The patient's sponta¬
neous speech demonstrates only a mini¬
mal abnormality—occasional wordfinding problems—that might be
attributed to an impairment of the mo¬
tor programmer. Additionally, this hy¬

pothesis provides no principled account

ofthe dissociation between the retrieval
of phonologically, compared with se¬
mantically, guided retrieval. Finally, a
deficit with phonologically cued stimuli
was documented on the levels of pro¬
cessing task in which no verbal output
was necessary.
A second issue concerns the marked

dissociation between the patient's nor¬
mal phonological processing in natural¬
istic, "automatic" language tasks such
as spontaneous speech, reading, and
naming and his marked impairment on
memory and fluency tasks requiring the
strategically guided, "controlled" re¬
trieval of phonologically specified
items. Following Schneider and Shiffrin41 and Shiffrin and Schneider42, a con¬
trolled process refers to an attentionrequiring procedure that may be
accessed or developed for the comple¬
tion of a novel or unusual task. Tasks
such as the production of lists of rhym¬
ing words that are not part of normal
language routines may require a strate¬
gically guided or controlled search of
lexical phonology. An automatic pro¬
cess, in contrast, is effortless and pro¬
ceeds without active control; the seman¬
tically guided activation of lexical
phonology that is thought to character¬
ize normal routines involved in speech
production43 and object naming may in¬
volve an automatic process.
Within this framework, one possible
explanation for the patient's deficits is
an impairment in the controlled search
of stored information. This hypothesis
may take at least two forms. First, one
might propose that the deficit is specific
to the domain of lexical phonology; in
this hypothesis, the controlled search of
lexical semantics is assumed to be nor¬
mal. Because tasks such as the retrieval
of phonology on the basis of rhyme or
initial letter are assumed to require con¬
trolled search of phonological informa¬
tion, whereas spontaneous speech,
naming repetition, and reading are as¬
sumed to be mediated by the relatively
automatic procedures involved in nor¬
mal speech production, this hypothesis
for the
can
account
patient's

performance.
A second explanation is that the pa¬
tient suffers from a general impairment
in controlled search. On this hypothesis,
the difference in performance between
phonologically and semantically speci¬
fied items might reflect differences in
the amount of effort required for con¬
trolled searches of the semantic and
phonological stores, which, in turn, may
reflect discrepancies in the organization
of these stores; an impaired procedure
for controlled search may be more effec¬
tive operating within the semantic do¬
main, in which related items may be

"linked," compared with phonological

stores, in which such interrelationships
are weak or absent. One

might expect,

for example, that the retrieval of animal
names would be facilitated by the multi¬
plicity of semantic and, perhaps, visual
attributes that they share; in contrast,
preexisting linkages between words

Downloaded From: http://archneur.jamanetwork.com/ by a Michigan State University User on 06/11/2015

sharing an initial letter in the phonologi¬
cal lexicon would be expected to be
weak.
The

hypothesis that the patient's

symptoms reflect an impairment in con¬

trolled search of stored information also
accounts for another interesting aspect
of his performance: the relatively nor¬
mal recency effect, but the failure to
demonstrate primacy. Recency effects
are typically attributed to the mainte¬
nance of phonological information in the

articulatory loop or phonemic buffer;
primacy effects, in contrast, are typical¬
ly attributed to the activation of lexical
information. Thus, the ability to pro¬
duce words from the beginning of a list
(ie, primacy) may require the controlled
search of lexical phonology, which, in
our hypothesis, is assumed to be
impaired.
Anatomic studies in monkeys provide
data consistent with the proposal that
the patient's dorsolateral frontal lobe

infarct is associated with a defect in the
controlled search of lexical phonology.
Chavis and Pandya,44 for example, dem¬
onstrated that the frontal lobes (includ¬
ing the dorsolateral, medial, and orbito¬
frontal portions) have reciprocal
connections with the primary auditory
and auditory association cortex. One
might speculate that the left dorsolater¬
al frontal lobe plays a critical role in the
retrieval of lexical phonology by virtue
of its role in the activation of the audi¬
tory association cortex (including Wer¬
nicke's area), which may support lexical
phonology in man.
An important, but thus far unan¬
swered, question is whether the puta¬
tive deficit in controlled search under¬
lies the patient's memory impairment.

Although one might expect a general

deficit in controlled search of stored in¬
formation to cause a memory impair¬
ment, in the absence of an explicit ac¬
count of the role of phonological factors
in encoding and retrieval, it is not clear
whether a specific impairment in the
retrieval of phonologically specified in¬
formation would cause amnesia. It is
also possible, of course, that the impair¬
ment in controlled search and memory
deficit are unrelated but co-occur be¬
cause both processes are dependent on
structures in the left dorsolateral fron¬
tal lobe.
An additional issue concerns the im¬
plications of our patient's performance
for conceptions regarding the relation¬
ship between the frontal lobe and mem¬
ory.

Investigations involving

mon¬

keys45,46 and man47,48 have suggested that
frontal lobe damage may be associated
with specific memory impairments.
Other researchers, however, have ar¬
gued that the apparent memory failure

observed with frontal lobe damage does

not represent a specific impairment of

encoding or retrieval but a disruption of
more complex forms of human behav¬
ior.9,29 Luria,49 for example, argued that
patients with frontal lobe damage are

unable to maintain a stable intention to
remember and are plagued by "patho¬
logic inertia" and "high distractibility."
In a similar vein, Warrington and
Weiskranz34 proposed that the frontal
lobes play a critical role in the cognitive
mediation mechanism, which serves to
organize, manipulate, and embellish
memory input. Signoret and Lher¬
mitte0" provided empirical support for
the proposal that memory deficits asso¬
ciated with frontal lobe damage may be
related to a deficit in the organization of
material to be learned, rather than en¬
coding or retrieval per se; these investi¬
gators described two patients with fron¬
tal lobe damage who were severely

impaired in paired associate list learn¬
ing but were relatively unimpaired
when the stimuli were presented in a
sentence context or with instructions to

generate a visual image of the named

objects.
Although the data derived our pa¬
tient confirm previous observations
that even relatively minor frontal lobe
damage may be associated with a pro¬
found memory impairment, these data
do not permit one to adjudicate between
these possibilities. To the extent that
the memory impairment is attributable
to the putative deficit in controlled
search of lexical phonology, the data
support the thesis that frontal lobe am¬
nesia is attributable to higher-level im¬
pairments rather than to impaired en¬
coding or retrieval. Alternatively, ifthe
memory deficit and impairment in con¬
trolled search are unrelated, the data
may be consistent with the view that the

frontal lobes are critical for encoding
and/or retrieval.
Finally, one might wonder why the at
least relatively selective impairment in
processing of phonologically cued infor¬
mation demonstrated by our patient has
not been reported previously. One pos¬
sibility is that the impairment has been
confounded by the frequent association
of dorsolateral frontal lobe lesions with
nonfluent aphasia; impairments in the
motor aspects of speech production or
lexical phonology, which are often ob¬
served in association with left dorsolat¬
eral frontal lobe infarcts, might obscure
the effects of impairment in controlled
search.
This study was supported, in part, by the Medical
Research Service of the Department of Veterans
Affairs and the Memory Disorders Clinic of the
Division of Aging, Department of Health and Reha¬
bilitative Services of the State of Florida,
Gainesville.

References
1. Risse GL, Rubens AB, Jordan LS. Disturbances of long-term memory in aphasic patients.
Brain. 1984;107:605-617.
2. DeRenzi E, Nichelli P. Verbal and nonverbal
short-term memory impairment following hemispheric damage. Cortex. 1975;11:341-354.
3. Cermak LS, Butters N, Moreines J. Some
analyses of the verbal encoding deficit of alcoholic
Korsakoff patients. Brain Lang. 1974;1:141-150.
4. Cermak LS, Moreines J. Verbal retention deficits in aphasic and amnesic patients. Brain Lang.

1976;3:16-27.

5. Cermak LS, Stiassny D, Uhly B. Reconstructive retrieval deficits in Broca's aphasia. Brain

Lang. 1984;21:95-104.
6. Mohr J, Pessin M, Finkelstein S, Funkenstein
H, Duncan G, Davis G. Broca's aphasia: pathologic
and clinical aspects. Neurology. 1978;28:311-324.
7. Foix C. Aphasies. In: Roger G, Widal P, Tessier PJ, eds. Nouveau Traite de Medicine. New
York, NY: Masson Publishers USA Inc; 1928:223\x=req-\
246.
8. Milner B. Some effects of frontal lobectomy in
man. In: Warren JM, Akert K, eds. The Frontal
Granular Cortex and Behavior. New York, NY:
McGraw-Hill International Book Co; 1964:313-334.
9. Benton AL. Differential behavioral effects in
frontal lobe disease. Neuropsychologia. 1968;6:53\x=req-\
60.
10. Dunn LM. Peabody Picture Vocabulary
Test. Circle Pines, Minn: American Guidance Service; 1959.
11. Kertesz A. Aphasia and Related Disorders.
New York, NY: Grune & Stratton; 1979.
12. Goodglass H, Kaplan E. Assessment of
Aphasia and Related Disorders. Philadelphia, Pa:
Lea & Febiger; 1982.
13. DeRenzi E, Vignolo LA. The token test: a
sensitive test to detect receptive disturbances in
aphasia. Brain. 1962;85:665-678.
14. Kaplan E, Goodglass H, Weintraub S. Boston Naming Test. 1976. Experimental edition.
15. Heilman KM, Scholes R, Watson RT. Defects of immediate memory in Broca's and conduction aphasia. Brain Lang. 1976;3:201-208.
16. Gonzalez Rothi L, Coslett HB, Heilman KM.
Battery of Adult Reading Function. 1983. Experimental edition.
17. Goodglass H, Gleason JB, Bernholtz NA,
Hyde MR. Some linguistic structures in the speech
of a Broca's aphasic. Cortex. 1972;8:191-212.
18. Berg EA. A simple objective test for measuring flexibility in thinking. J Gen Psychol.

1948;39:15-22.

19. Stroop JR. Studies of interference in serial
verbal reactions. J Exp Psychol. 1935;18:643-662.
20. Christensen A. Luria's Neuropsychological
Investigations. Holliswood, NY: Spectrum Publications; 1981.
21. Wechsler D. A standardized memory scale
for clinical use. J Psychol. 1945;19:87-95.
22. Taylor EM. The Appraisal of Children with
Cerebral Deficits. Cambridge, Mass: Harvard University Press; 1959.
23. Milner B. Visual recognition and recall after
right-temporal lobe excision in man. Neuropsycho-

logia. 1968;6:191-209.
24. Kimura D. Right temporal lobe damage.

Arch Neurol. 1963;8:48-56.
25. Huppert F, Piercy M. Recognition memory
in amnesic patients: effect of temporal context and
familiarity of material. Cortex. 1976;12:3-20.
26. Squire L. Two forms of human amnesia: an
analysis of forgetting. J Neurosci. 1981;1:635-640.
27. Freed D, Corkin S, Cohen N. Forgetting in
H.M.: a second look.
Neuropsychologia.

1987;25:461-471.

28. Peterson LR, Peterson MJ. Short-term retention of individual verbal items. J Exp Psychol.

1959;58:193-198.
29. Butters N, Grady G. Effect of predistractor
delays on the short-term memory performance of
patients with Korsakoff's and Huntington's disease. Neuropsychologia. 1977;15:701-706.
30. Wickens DD. Encoding categories of words:
empirical approach to meaning. Psychol Rev.

an

1970;77:1-15.

31. Moscovitch M. Multiple dissociations offunction in amnesia. In: Cermak L, ed. Human Memory
and Amnesia. Hillsdale, NJ: Lawrence Erlbaum;
1982:337-370.
32. Freedman M, Cermak LS. Semantic encoding deficits in frontal lobe disease and amnesia.
Brain and Cognition. 1986;5:108-114.
33. Craik FIM, Lockhart RS. Levels of processing. A framework for memory research. J Verbal
Learn Verbal Behav. 1972;11:671-684.
34. Warrington EK, Weiskranz L. Amnesia: a
disconnection
syndrome? Neuropsychologia.

1982;20:233-248.
35. Benton AL, Hamsher KD. Multilingual

Aphasia Examination. Iowa City, Iowa: University of Iowa; 1976.
36. Jones-Gotman M, Milner B. Design fluency:
the investion of nonsense drawings after focal cortical lesions. Neuropsychologia. 1977;15:653-674.

Downloaded From: http://archneur.jamanetwork.com/ by a Michigan State University User on 06/11/2015

37. Caramazza A, Berndt RS, Basili AG. The
selective impairment of phonological processing: a
case study. Brain Lang. 1983;18:128-174.
38. Shallice T, Warrington EK. Auditory-verbal
short-term memory impairment and conduction
aphasia. Brain Lang. 1977;4:479-491.
39. Caramazza A, Basili AG, Koller JJ, Berndt
RS. An investigation of repetition and language
processing in a case of conduction aphasia. Brain

Lang. 1981;14:235-271.
40. Martin RC. Articulatory and phonological

deficits in short-term memory and their relation to
syntactic processing. Brain Lang. 1987;32:159-192.
41. Schneider W, Shiffrin RM. Controlled and
automatic human information processing, I: detection, search and attention. Psychol Rev. 1977;84:1\x=req-\
66.
42. Shiffrin RM, Schneider W. Controlled and
automatic human information processing, II: Perceptual learning, automatic attending and a general
theory. Psychol Rev. 1977;84:127-189.
43. Garrett M. Levels of processing in sentence
production. In:Butterworth B, ed. Language Production. London, England: Academic Press;
1980:177-220.
44. Chavis DA, Pandya DN. Further observations on corticofrontal connections in the rheusus
monkey. Brain Res. 1976;117:369-386.
45. Gross CG, Weiskranz L. Some changes in
behavior produced by lateral frontal lesions in the
macaque. In: Warren JM, Akert K, eds. The Frontal Granular Cortex and Behavior. New York,
NY:McGraw-Hill International Book Co; 1964:74\x=req-\
101.
46. Mishkin M, Vest B, Waxier M, Rosvold HE.
A re-examination of the effects of frontal lesions on
object alternation. Neuropsychologia. 1969;7:357\x=req-\
363.
47. Alexander M, Freedman M. Amnesia after
anterior communicating artery aneurysm rupture.

Neurology. 1984;34:752-757.
48. Volpe B, Hirst W. Amnesia following rupture and repair of an anterior communicating artery aneurysm. J Neurol Neurosurg Psychiatry.
1983;46:704-709.
49. Luria A. Memory disturbance in local brain
lesions. Neuropsychologia. 1971;9:367-375.
50. Signoret J, Lhermitte F. Amnesic syn-

dromes and the encoding process. In: Rosenzweig
MR, Bennett EL, eds. Neural Mechanisms of
Learning and Memory. Cambridge, Mass: MIT
Press; 1976:67-75.