Associations between Phenylalanine-to-Tyrosine Ratios and Performance on Tests of Neuropsychological Function in Adolescents Treated Early and Continuously for Phenylketonuria

by Monica Luciana, Jill Sullivan, Charles A. Nelson
Associations between Phenylalanine-to-Tyrosine Ratios and Performance on Tests of Neuropsychological Function in Adolescents Treated Early and Continuously for Phenylketonuria
Monica Luciana, Jill Sullivan, Charles A. Nelson
Child Development
Start Page: 
End Page: 
Select license: 
Select License

Child Development, November/December 2001, Volume 72, Number 6, Pages 1637-1652

Associations between Phenylalanine-to-Tyrosine Ratios
and Performance on Tests of Neuropsychological Function
in Adolescents Treated Early and Continuously
for Phenylketonuria

Monica Luciana, Jill Sullivan, and Charles A. Nelson

Phenylketonuria (PKU) is a genetic disorder characterized by hyperphenylalaninemia. Treatment involves dietary phenylalanine restriction to prevent mental retardation. Because phenylalanine is involved in tyrosine synthesis and tyrosine is a catecholamine precursor, low tyrosine may lead to brain dopamine deficiencies. Because dopamine is involved in the modulation of prefrontally orchestrated executive functions, deficiencies may lead to executive impairments. Despite treatment, impairments in executive cognitive functions have been reported in young children with PKU. Outcome beyond middle childhood has not been extensively investigated. In this study, PKU-affected adolescents (N = 18)with normal-range IQ scores completed neuro- psychologicaI tests, and their performance was compared with unaffected peers (N = 16) and chronically ill controls (N = 17).Results demonstrated that the overall performance of the PKU group did not differ from that of the other two groups, but that performance of the PKU proband was associated with phenylalanine and tyrosine levels, and most strongly with phenylalanine-to-tyrosine ratios at several points in development. These findings provide a preliminary test of the dopamine hypothesis of PKU as it applies to adolescents and

young adults.


Phenylketonuria (PKU) is an inborn disorder of metab- olism affecting approximately 1 in 10,000 to 20,000 live births (Benson & Fensom, 1985). PKU is caused by a limited ability to metabolize the amino acid phenyl- alanine due to the absence or inactivity of the enzyme phenylalanine hydroxylase (PAH). Left untreated, af- fected individuals experience a rise in serum phenyl- alanine and subsequent mental retardation. The biochemical mechanisms that explain why high phenylalanine levels result in mental retardation have not been definitively established. However, under normal circumstances, phenylalanine, through the ac- tion of PAH, is converted to tyrosine, the precursor enzyme for the production of dopamine and norepi- nephrine in the central nervous system (McKean & Peterson, 1970). Treated individuals are typically maintained on a diet in which phenylalanine is re- stricted, but not necessarily entirely eliminated (due to the body's need for protein), and tyrosine is sup- plemented. Nonetheless, a state of brain tyrosine (and dopamine) depletion is still possible due to (1)the shared competitive mechanism through which both phenylalanine and tyrosine cross the blood-brain barrier, and (2) the differential sensitivity of region-


ally specific brain neurons to modest decreases in tyro- sine availability (for a thorough review, see Diamond, Prevor, Callender, & Druin, 1997). Thus, even with treatment, phenylalanine levels in PKU may be ele- vated in the context of low tyrosine levels. It has been suggested (Diamond et al., 1997) that because of this biochemical state, characterized by moderately high phenylalanine levels in the context of low tyrosine levels, one of the best indicators of dopamine avail- ability in PKU is the phenylalanine-to-tyrosine ratio (Phe:Tyr). Ths ratio is not typically examined in studies of PKU.

Dopamine neurons in the prefrontal cortex and ret- ina, unlike those in other regions, are particularly sen- sitive to decreased tyrosine availability (Bradberry, Karasic, Deutch, & Roth, 1989; Diamond & Herzberg, 1996; Tam, Elsworth, Bradberry, & Roth, 1990). Re- duced levels of the precursor can theoretically occur either when PKU is left untreated, because phenyl- alanine cannot be converted to tyrosine through PAH activity, or when PKU is treated, if the ratio of phe- nylalanine to supplemented dietary tyrosine is in the moderate to high range. Under either condition, dis- ruptions in prefrontally mediated behaviors are pos- sible, although specific prefrontal dysfunction is most likely when phenylalanine levels are moderately, but not excessively, elevated. Diffuse damage, extending beyond prefrontal cortex, is more likely when phe-

02001 by the Society for Research in Child Development, Inc All rights reserved. 0009-3920/2001/7206-0003

nylalanine levels are excessively high, as in untreated PKU (Diamond et al., 1997).

Definitive evidence of central nervous system dopamine deficiency in PKU has not been demon- strated, although there is limited evidence of reduced tyrosine concentrations in the central nervous sys-tems of phenylketonurics (McKean & Petersen, 1970). Additionally, aberrant levels of dopamine precur- sors, metabolites, and reduced norepinephrine in the plasma, urine, and cerebrospinal fluid of individuals who have discontinued dietary treatment have been reported (Guttler & Lou, 1986). An inverse relation- ship between plasma phenylalanine and urinary dopamine excretion also has been found (Krause et al., 1985).

In laboratory animals, when dopamine is depleted prenatally through tyrosine depletion or neurochem- ical lesions, animals exhibit decreased responding to dopamine-dependent rewarding stimuli, as well as impaired acquisition of operant responding (Moy, 1995; Palmour, Ervin, Baker, & Young, 1998; Yoko- goshi & Nomura, 1991). Supersensitivity to dopamin- ergic agents has also been observed (Owasoyo, Neri, & Lamberth, 1992). Moreover, when tyrosine or other dopamine agonists are administered prenatally to pregnant females, their offspring show lifetime alter- ations in response to dopaminergic agents (Nasello & Ramirez, 1978; Santana, Martin, & Rodrigues, 1994). Acute tyrosine depletions in healthy adult humans have been associated with increases in arousal, de- pression, and anxiety (McCann et al., 1995), as well as impaired cognitive function (McCann et al., 1992) and decreased responses to the rewarding effects of alcohol (Rammsayer & Vogel, 1995). Alternatively, under stressful conditions, tyrosine administration improves learning and worlung memory performance in animals (Ahlers, Salander, Shurtleff, & Thomas, 1992; Shukitt-Hale, Stillman, & Lieberman, 1996; Shurtleff, Thomas, Ahlers, & Schrot, 1993) and in humans (Deijen & Orlebeke, 1994; Shurtleff, Thomas, & Schrot, 1994).

Although experimental studies of tyrosine depletion have yet to result in an animal analogue of human mental retardation, it is well established that dopamine modulates cognitive functions, particularly those me- diated by the prefrontal cortex. Specifically, dopamine in the dorsolateral prefrontal cortex facilitates spatial working memory performance (Diamond, 1996; Luci- ana & Collins, 1997; Sawaguchi & Goldman-Rakic, 1991). Nonhuman primates with depleted prefrontal dopamine exemplify impairments in working mem- ory, and dopamine augmentation under acute condi- tions improves performance (Brozoski, Brown, Ros- vold, & Goldman, 1979).

Thus, although mental retardation as a consequence of PKU has been successfully averted through dietary intervention (Williamson, Koch, Azen, &Chang, 1981), it has yet to be determined whether phenylke- tonurics exemplify dopamine-dependent cognitive deficits. Although language, memory, and basic mo- tor skills appear to be normal, several researchers have reported deficits in reaction times, attention, problem solving, abstract reasoning, and executive abilities in affected groups of all ages (Waisbren, Brown, de Sonne- ville, & Levy, 1994). Executive functions involve the ability to maintain a problem-solving set that is ap- propriate for goal acquisition, including formulation, planning, self-monitoring, mental flexibility, and the ability to change strategies in response to new infor- mation. Tests of executive functioning typically re- quire information to be held in mind while other pieces of information are simultaneously processed (e.g., multitashg or working memory). Several studies have reported that infants and young children with early treated PKU show deficits in working memory and other high-level reasoning skills, despite the fact that these children have normal-range IQs (Diamond et al., 1997; Pennington, van Doorninck, McCabe, & McCabe, 1985; Spreen, Tupper, & Risser, 1984; Welsh, Pennington, Ozonoff, Rouse, & McCabe, 1990). Out- come in young adults treated since birth has not been reliably ascertained, because findings have been in- consistent in studies of older patients. Mazzocco et al. (1994) studied PKU probands, ages 6 to 13 years, and found no evidence of prefrontal dysfunction, suggest- ing that the cognitive deficits observed in younger children represent a developmental delay. In con- trast, Weglage, Pietsch, Funders, Koch, and Ullrich (1996) found evidence of dysfunction in PKU subjects aged 8 to 13 years on the Stroop test, a conventionally used executive measure of attention and inhibitory control. Ris and colleagues (Ris, Williams, Hunt, Berry, & Leslie, 1994) compared the neuropsycholog- ical performance of adults who had PKU with that of their unaffected siblings and found that the PKU probands were relatively impaired on measures of at- tention, visual reasoning, and on the Wisconsin Card Sort. Smith and colleagues (Smith, Klim, Mallozzi, & Hanley, 1996) studied the neuropsychological perfor- mance of hyperphenylalaninemic adults and found that the hyperphenylalaninemic group performed worse than control subjects on both frontal and tem- porallparietal lobe tests. Only performance on the frontal measures, however, was related to concurrently measured phenylalanine levels.

Interpretation of these studies is complicated, as has been noted (Stemerdink et al., 1999), because the neuropsychological measures vary widely between studies, and several of the available reports (e.g., Ris et al., 1994; Smith et al., 1996) focused on samples of young adult PKU probands who had been early but not continuously treated. Most recently, Stemerdink et al. (1999) assessed neuropsychological function in early and continuously treated PKU probands, ages 8 to 20, relative to an age-matched control group. The cognitive measures utilized involved a test of visual contrast sensitivity, the Wisconsin Card Sort, the Tower of Hanoi planning test, and the Corsi-Milner test of recency judgments. Several control measures were also employed to assess nonexecutive functions. These authors reported that their PKU sample was impaired relative to normal controls on a composite index of the prefrontal measures and specifically on the Wisconsin Card Sort and the test of recency judg- ments. Performance on nonexecutive measures was indistinct between the groups. Moreover, concurrent levels were related to task performance in the pre- dicted direction within the youngest (e.g., age 8 years) and oldest (age 20 years) subjects, although findings were not consistent across the tasks em- ployed. In the intermediate age groups, concurrent phenylalanine levels were related to task performance in a direction opposite to that predicted. That is, high concurrent phenylalanine levels were associated with better task performance. There have been several re- ports of associations between concurrent phenylala- nine levels and composite indices of prefrontal func- tion (Diamond et al., 1997; Ris et al., 1994; Weglage et al., 1996; Welsh et al., 1990), although the relationship between phenylalanine levels and performance on specific tests of executive function has been inconsis- tent. None of these studies of adolescent and young adult probands has examined the relationship be- tween Phe:Tyr ratios and neuropsychological perfor- mance, despite its potential importance for the study of dopamine availability in PKU.

This literature raises several questions relevant to the longer term outcome of adolescents and young adults with PKU. Not only does it remain uncertain as to whether early and continuous restriction of phe- nylalanine prevents cognitive dysfunction, particularly with respect to behaviors mediated by the prefrontal cortex, but evidence for a dopaminergic mechanism underlying the problems that have been reported is lacking.

In this study, we compared the neurocognitive per- formance of adolescents and young adults who had early and continuously treated PKU with that of two comparison groups-one consisting of age-matched peer-nominated controls and the other an age-matched group of individuals with chronic illnesses. All indi- viduals completed measures of prefrontally mediated

Luciana, Sullivan, and Nelson 1639

executive function including spatial working memory, set shifting and look-ahead planning. The measures of executive function used in this study are sensitive to al- terations in catecholamine neurotransmission (Coull, Middleton, Robbins, & Sahaluan, 1995; Dias, Robbins, & Roberts, 1996; Downes et al., 1989; Lange et al., 1992; Morris et al., 1988) and have been used to detect cognitive impairments in developmentally disordered populations (Hughes, Russell, & Robbins, 1994; Luciana, Lindeke, Georgieff, Mills, & Nelson, 1999; Ozonoff, 2001). It was predicted that relative to con- trols, individuals with PKU would show impair- ments in executive functions such as spatial working memory and set shifting that are modulated by dopamine activity in the frontal cortex. It was also predicted that individual differences in the cognitive performance of PKU probands on these tests of exec- utive function would be associated with the Phe:Tyr ratio measured at various time points during individ- uals' lifetimes. A high ratio score, indicating low ty- rosine in the context of high phenylalanine levels, was expected to be associated with relatively poor cognitive performance.


PKU probands were recruited from the patient da- tabase at the Fairview-University Medical Center's Department of Pediatrics at the University of Minne- sota. All had received treatment at the hospital's PKU Clinic during childhood, had been diagnosed with classic PKU (i.e., based on the clinic's criteria of serum phenylalanine concentrations >20 mgldl on an unre- stricted diet, normal or low tyrosine levels, and in- creased phenylketones present in urine), and had begun treatment within the first 3 months of life. This clinic serves Minnesota and several surrounding states, as well as southern central Canada. The clinic database was screened for individuals between the ages of 14 and 25 who had measured full-scale IQ scores above 80 on the age-appropriate version of the Wechsler Intelligence Scales. A total of 58 probands were identified as possible study participants, and all were sent letters of invitation.

Letters were followed by phone calls from a study representative. Recruitment was complicated by the fact that many families were currently living out of state or in Canada. Of the initial pool of 58 possible participants, 19 could not be contacted, 11 refused to participate, and 28 expressed some interest in the study. Of those who expressed interest, 10 partici- pants initially agreed to participate but did not make their scheduled appointments even when multiple visits were scheduled. The final number of PKU probands who agreed to participate (N = 18) were asked (but not required) to nominate a friend of the same age and gender to co-participate in the study. These peer controls (N = 16) were contacted sepa- rately and invited to participate in a study of adoles- cent health.

In an attempt to address whether the PKU pro- bands who had agreed to participate in this study were representative of the group as a whole, the IQ scores and phenylalanine levels of the 18 PKU probands who participated in this study were com- pared with those of 30 affected individuals who were identified as possible participants but could not be contacted or declined to participate. The nonpartici- pants had phenylalanine, tyrosine, and IQ data avail- able through medical chart review. Although scores remained within the normal range, there was a trend toward lower full-scale IQ scores in the nonpartici- pating (M = 97.2, SD = 11.3) versus participating (M = 104.8, SD = 10.9) PKU sample, F(1,37) = 2.76, p = .lo. Phenylalanine levels did not differ between the groups, F(1, 37) = 1.14, ns.

In order to control for the nonspecific effects of chronic illness on behavioral functioning, a group of age-matched controls with lifelong illnesses was also recruited. This comparison group was comprised of patients with cystic fibrosis (n = 6) and juvenile onset diabetes (n = 4) identified through records of the Pediatric Pulmonary and the Pediatric Endocrinol- ogy Clinics at the Fairview-University Medical Cen- ter, as well as asthmatic patients (n = 7) from the Pul- monology Clinic at Children's Health Care in St. Paul, MN. Of 95 potentially recruitable chronically ill patients, 73% (N = 32) of patients with cystic fibro- sis, 73% (N = 22) of patients with diabetes, and 64% (N = 14) of patients with asthma declined to partici- pate. Eleven patients agreed to participate but either cancelled or failed to attend their scheduled visits. The 17 individuals who agreed to participate were also invited to nominate a peer to participate with them. There was no difference between the PKU and chronic illness groups in their willingness to recruit peers, x2(1, N = 37) = 2.45, p = 12. Participants under the age of 18 years (65%) were significantly more likely to re- cruit peers that were older participants, x2(1, N = 37) = 8.40, p < .01.

The final study sample consisted of 18 PKU probands, 16 peer controls, and 17 patients with other chronic illnesses. All individuals were assessed on an outpatient basis. Demographic characteristics of the groups are presented in Table 1. As indicated, the three samples were comparable in terms of gender repre- sentation, age, F(2, 51) = 1.61, ns, ethnicity, dwelling (rural versus urban settings), marital status, and liv- ing arrangements. As would be expected, age at ini- tial diagnosis was significantly different between PKU probands (M = 23.65 days, SD = 18.58 days) and chronically ill participants (M = 4.74 years, SD =

4.13 years). All but one chronically ill participant was currently participating in daily medical and/or di- etary treatment; the majority of these participants (N = 13) reported that they complied with treatment as prescribed at least 75% of the time, with no significant differences in rates of compliance between the chron- ically ill subgroups, x2(2, N = 17)= 5.42, p = .49. Socio- economically, the groups were similar in terms of their completed years of education, F(2,51) = 1.18, ns, level of education of their parents, and their annual median incomes (their own if living independently; parental if living with parents). The three groups did not differ in terms of history of academic difficulties, F(2,51) = 1.05, ns, the number of individuals who re- peated a grade in school, F(2, 51) = .26, ns, nor in the number of current clinical scale elevations on the Minnesota Multiphasic Personality Inventory- Adolescent Version (MMPI-A; Butcher et al., 1992), F(2, 51) = .64, ns. There was a group difference in the number of individuals who had received past psychi- atric treatment, with approximately equal numbers of PKU and chronically ill individuals reporting such histories. No one in the peer control group reported a psychatric treatment history, F(2,51) = 6.31, p < .01.

Cognitive Assessment

Tests of Executive Function

Three tasks were selected to represent executive functions mediated by the prefrontal cortex. These in- cluded a test of spatial working memory (Owen, Downes, Sahakian, Polkey, & Robbins, 1990), the Tower of London planning task (Shallice, 1982), and a set-shifting task, analogous to the Wisconsin Card Sort (Downes et al., 1989). All tasks were adminis- tered through the use of a touch-screen computer (Fray, Robbins, & Sahaluan, 1996).

Spatial working memory. This self-ordered searching task measures worlung memory for spatial stimuli and requires the participant to use mnemonic infor- mation to work toward a goal (Owen et al., 1990; Owen, Evans, & Petrides, 1996; Petrides & Milner, 1982). Two variables of interest are computed. The first is a measure of strategy whereby a high score represents poor strategy use. The second consists of mnemonic errors for searches of two, three, four, six, and eight items.

Luciana, Sullivan, and Nelson 1641

Table 1 Demographic Characteristics of Phenylketonuria (PKU) Probands, Peer Controls, and Chronically I11

  PKU Probands Peer Controls Chronically I11 Patients
Age Gender (ma1e:female) Ethnicity (% White) Years of education      
Number with reported history of academic difficulties      
Number who repeated a grade in school Number with history of psychiatric treatment Average number of MMPI clinical scale elevations      
Dwelling Urban      
Marital status      
Single Married      
Living arrangements With parents or guardian Dormitory Rented apartment Homeowner      
Employment status Student      
Student, plus part-time employment Part-time employment Full-time employment Unemployed      
Mothers' education      
<12 years 12 yearsa or GED Some collegeb College degree Advanced degree Unknown      
Fathers' education      
<12 years 12 yearsa or GED Some collegeb College degree Advanced degree Unknown      


Note: MMPI = Minnesota Multiphasic Personality Inventory; GED = Graduate Equivalency Diploma

a Indicates completion of high school degree. Includes trade school. Participants reported personal income if financially dependent, otherwise parental income was used. Two participants refused to provide financial information.

Infradimensional/extradimensionalset shifting. One performance is typically attributed to the inability to of the most popularly used measures for the experi- shift response set between conceptual categories. mental assessment of frontal lobe dysfunction is the However, the precise nature of the cognitive deficit

Wisconsin Card Sort Test (Milner, 1964). Deficient underlying task failure has yet to be determined, be- cause the task actually requires several abilities. To dissociate the various abilities underlying set-shift- ing performance, we utilized a task called the "In- tradimensional /Extradimensional Set-Shift task (Dias, Robbins, & Roberts, 1996; Downes et al., 1989). This task measures discrimination and reversal learning under conditions that require the subject to shift at- tention to changing patterns of visual stimuli. Briefly, this task progresses along a series of nine stages of in- creasing difficulty. The first two stages involve sim- ple discrimination and reversal learning. The next three stages involve extending the discrimination and reversal learning to instances in which a distrac- tor stimulus is present. The sixth stage demands an attentional shift. Termed the intradimensional (ID) shift stage, novel or never-seen exemplars of each of the two prior dimensions are introduced, and the participant generalizes the rule from previous learn- ing in order to achieve correct responses. Subse- quently, another extradimensional (ED) attentional shift is required. Novel exemplars of each stimulus dimension are presented, and the subject must shift response set from the previously relevant dimension to the previously irrelevant dimension. This stage is analogous to the between-category shifts required by the Wisconsin Card Sort (Milner, 1964). Vari- ables coded for each participant included the stage reached, the trials to criterion, and number of errors for each completed stage.

Tower of London. This task measures planning and behavioral inhibition (Shallice, 1982). The participant views two displays on a computer screen, one of which is the target display and one that comprises a workspace. The participant must move colored balls on the screen to make the workspace display appear like the target. There are a minimum number of moves in which to complete a trial perfectly, and the participant is told at the start of a trial that the trial should be completed in that minimum number of moves. The following variables were computed for each problem set representing four levels of problem difficulty (2-move, 3-move, 4-move, and 5-move prob- lems): average number of moves to complete each set, planning time, and total number of problems perfectly completed in the minimum number of moves.

Performance expectationsfor PKU probands. Relative to comparison groups, we anticipated the perfor- mance of PKU probands on measures of executive function to be consistent with frontal lobe impair- ment. Specifically, the PKU group was expected to demonstrate increased mnemonic errors and higher strategy scores (indicative of poor use of strategy) on the spatial working memory task, an inability to pro- ceed through all nine stages of the IDIED set-shift task, and fewer perfect solutions in completing Tower of London problems.

Tests of Nonexecutive Function

Three additional cognitive measures were included to represent nonexecutive functions related to psycho- motor speed and accuracy, sequencing ability, and recognition memory.

Motor screening task. This test measures psychomotor speed and accuracy. The participant's task is to quickly and accurately touch visual targets that are presented one at a time on the computer screen. Ac- curacy and response latency are recorded.

Spatial span task. This task measures memory for a figural sequence and is a computerized analog of the Corsi block task (Milner, 1971). It yields a measure of the participant's nonverbal memory span, the num- ber of items that can be accurately remembered in a sequence.

Patternlspatial recognition. This delayed-match-to- sample task is a measure of recognition memory for visual patterns and spatial locations. First, the partic- ipant views geometric patterns presented consecu- tively on screen. After a brief pause, the participant is presented with two geometric patterns. One of the two designs is from the previously viewed list, while the other is novel. The participant indicates the pat- tern that has already been seen. Accuracy and re- sponse latency are recorded. An additional series of trials comprises a delayed-match-to-location task. The participant attends to the locations of five empty boxes presented at various locations on the screen. Then the participant views two empty boxes, one of which is in a location that was just targeted, and touches the box that represents a previously targeted location. Nota- bly, performance on these recognition memory tasks is highly sensitive to temporal and parietal lobe le- sions in adult patients and appears to be unaffected by frontal lobe pathology (B. J. Sahakian, personal communication, April 16, 1998).

Participants completed cognitive testing in the course of one 3-hour testing session. In addition, each participant completed the age-appropriate version of the MMPI-2 /A (Butcher et al., 1992), the Tennessee Self- Concept Scale (Fitts & Warren, 1996), and a structured clinical interview (the Diagnostic Interview for Chil- dren and Adolescents-Revised-Adolescent Version for DSM-IV: Reich, Leacock, & Shanfeld, 1995; or the Structured Clinical Interview for DSM-IV, Non-Patient Version: First, Spitzer, Gibbon, & Williams, 1994), Cog- nitive testing took place during either the first or sec- ond hour of the assessment. The cognitive tasks were administered in the same order for all participants.

Chart review of medical records maintained by the Department of Pediatrics and the Biomedical Genetics Laboratory were utilized to obtain IQ and phenylala- nine and tyrosine levels for PKU probands. IQ scores for PKU probands were obtained from each partici- pant's most recent neuropsychological assessment (N= 3, Wechsler Adult Intelligence Scale-Revised; N = 15, Wechsler Intelligence Scale for Children-Revised, 3rd edition). The average IQ obtained (M = 104.8, SD = 10.9) was not significantly different from the nor- mal population mean (M = 100, SD = 15). IQ was not measured in the comparison groups. Additionally, each participant's lifetime history of recorded phenyl- alanine and tyrosine levels was recorded and used in subsequent data analyses. Over the course of their life- times, PKU probands each contributed an average of

58.67 assessments (M = 58.67, SD = 36.06, range = 1120) to this analysis. From each assessment, consisting of a phenylalanine and a tyrosine level, the following values were computed for each of the PKU probands:

  1. The most recently obtained phenylalanine and tyrosine levels obtained within 6 months prior to the cognitive assessment. All probands had values available.
  2. Yearly grand averages of phenylalanine and ty- rosine for each year of life; the grand average computed for the first year of life always ex- cluded the first value recorded subsequent to the proband's birth (e.g., the value from which the diagnosis of PKU was made). From these yearly averages, five variables were constructed separately for phenylalanine and tyrosine, rep- resenting grand averages for the years (a) 0 to 2 years, (b) 3 to 4 years, (c) 5 to 8 years, (d) 9 to 13 years, and (e) 14 to 15 years. The period encom- passing ages 16 to 18 was not included in the analysis because only a small percentage of the PKU sample (n = 10 or 55%) was old enough to contribute these values. These age ranges were selected based on developmental research using cognitive tasks identical or similar to the ones used in this study (Diamond et al., 1997; Luci- ana & Nelson, 1998).
  3. For each of these five time periods, the average Phe:Tyr ratio was computed.

This method of obtaining phenylalanine and ty- rosine levels from medical records has been used in other PKU studies (Diamond et al., 1997). To assess whether the most recently obtained levels were repre- sentative of each participant's lifetime pattern of di- etary control, correlations between these levels and lifetime averages were computed. These findings are presented in Table 2. As indicated in Table 2, the most

Luciana, Sullivan, and Nelson 1643

Table 2 Intercorrelations among Lifetime and Most Recent Phenylalanine (Phe) and Tyrosine (Tyr) Levels

Most Recent Most Recent Lifetime
Phe Level Tyr Level Phe Level
Most recent Tyr level -.502+ - -
Lifetime Phe level .840** -.600* -
Lifetime Tyr level -.453+ .703** -,338

Note: Values represent Pearson correlations. *p < .05; **p < .01; 'p< .lo.

recent and lifetime levels of phenylalanine and ty- rosine were significantly intercorrelated.


Data were analyzed using SPSS for Windows, version

10.0 (SPSS, 1999). Two primary questions were ad- dressed in the data analyses: (1) Do PKU probands dif- fer from the two comparison groups with respect to cognitive functions in general, and, specifically, with respect to executive functions? and (2) In individuals with PKU, what is the relationship between lifetime measures of phenylalanine and tyrosine levels and cognitive function?

Group Differences in Cognitive Function

Cognitive task variables were compared between groups using one-way analyses of variance (ANOVA). When group differences were found, post hoc com- parisons using Tukey's honestly significant difference were evaluated to determine the nature of significant group differences. As indicated in Table 3, the cogni- tive performance of adolescents with PKU was com- parable with that exhibited by the two age-matched comparison groups. There were no statistically signif- icant differences noted between groups on any cogni- tive task variables.

Following the methodology of Welsh et al. (1990), a composite score representing executive function (EF) was derived by converting raw scores on the frontal lobe measures to Z scores, using the total sample mean and standard deviation. After the Z score con- versions, the following variables, each of which is the best representative of frontal lobe dysfunction for the particular task, were averaged to obtain compos- ite EF scores: spatial working memory strategy score, number of minimum-move Tower of London solu- tions, and IDIED stage reached. Composites were computed so that a Z score above the mean reflected below-average performance.

As a discriminant measure, a second composite

Task Variable PKU Probands

Motor error Motor latency Memory span

Spatial working memory Easy-item errors Hard-item errors Strategy score

Tower of London planning Minimum move solutions Average moves to complete

3-move problems
4-move problems
5-move problems

Planning times 3-move problems 4-move problems 5-move problems

ID/ED set shifting Stage reached ID errors ID trials ED errors ED trials Reversal errors

Recognition memory Patterns: % correct Spatial: % correct Executive function composite Nonexecutive function composite

Note: ID = intradimensional; ED = extradimensional

score was similarly derived to represent performance on tasks measuring nonexecutive functions (NEF). This measure was derived from the following variables: motor screening accuracy, motor screening latency, spa- tial memory span, pattern recogrution percent correct performance, and spatial recognition percent cor- rect performance. The composite was scored such that a high score represented poor performance. EF and NEF composite scores were compared among the three groups in a one-way ANOVA yielding no signif- icant main effect for group for the EF nor for the NEF composite.

Because this pattern of findings was inconsistent with what has been observed in younger samples of PKU probands, we considered whether performance in the comparison groups accurately represented "normal" cognitive performance. Akin to the notion of assortative mating, it may be that individuals with PKU select friends who are similar in having some level of cognitive deficit. As indicated in Table 1 (items 4-8), the three groups did not differ in terms of

Peer Chronically Controls I11 Patients F Ratio p Value

average educational level, whether school services had ever been received in the past, or in the average number of clinical scale elevations obtained on the MMPI. A more thorough analysis of the emotional functioning of these groups has indicated that the three groups were generally healthy with respect to current psychopathology (Sullivan, in press). How- ever, each group consisted of some individuals who experienced difficulties in one or more of these areas, as indicated in Table 1.Accordingly, to further ad- dress the representativeness of the two comparison samples (peers and chronically ill), their cognitive performance was compared with that of a third group (n = 19) of age-matched normal controls whose per- formance on the same battery of tests was measured for a separate study. This sample was recruited from the undergraduate population of students at the Uni- versity of Minnesota and from a private area high school. Their data have been reported elsewhere (Lu- ciana & Nelson, 1998). No one in this additional sample had a history of school difficulty, special

education, chronic illness, or psychological disor- der. There were no significant differences in cognitive performance on any task variable among the peer- nominated controls, chronically ill adolescents, and this third normative sample of young adults. Thus, we have no reason to conclude that our comparison groups were not representative of individuals in this age range.

Associations between Biochemical Variables and Cognitive Outcome

Associations between phenylalanine levels, tyrosine levels, the Phe:Tyr ratio and measures of executive and nonexecutive function were examined. Several tables are included to describe these data: Table 4 presents correlations between biochemical variables at various points in development and the PKU probands' ages and IQs, Table 5 presents correlations between mea- sures of executive function and biochemical variables, and Table 6 presents correlations between measures of nonexecutive function and biochemical variables.

Phenylalanine and Cognitive Function

Of the 18 PKU probands studied, 77% (141 18) indi- cated by self-report that they were currently comply- ing with dietary restrictions. However, the range of most recently obtained phenylalanine levels varied in this group from 5.3 to 28.8 mg/dl (M = 16.3, SD = 6.8). Since adequate control of phenylalanine levels is represented by a value of less than 10 to 15 mg/ dl (Di- amond et al., 1997), many of these participants were either not maintaining the diet adequately or were not maximally benefiting from phenylalanine restriction.

Luciana, Sullivan, and Nelson 1645

The correlation between the most recently obtained phenylalanine level and IQ was in the predicted direc- tion, with high phenylalanine levels associated with lower full-scale IQ scores. Additionally, older partici- pants tended to exhibit higher phenylalanine levels (see Table 4). As indicated in Tables 5 and 6, there were significant associations between phenylalanine levels obtained at various points in development and several cognitive task variables.


With respect to spatial worlung memory performance, phenylalanine levels at 0 to 2 years of age were related to the total number of mne- monic errors, but in a manner opposite to that pre- dicted. High phenylalanine levels were associated with lower numbers of errors. However, poor use of strategy was associated with high phenylalanine lev- els at ages 3 to 5 years, 9 to 13 years, 14 to 15 years, and in the past 6 months. In terms of set-shifting per- formance, high phenylalanine levels at ages 3 to 4,5 to 8, and 9 to 13 years were associated with a decreased stage reached on the task. Tower of London perfor- mance, as assessed by a high number of items solved in the minimum number of moves, was associated with higher phenylalanine levels at 0 to 2 years, a find- ing similar to what was found for spatial working memory errors and counter to prediction. High phe- nylalanine levels at 9 to 13 years of age were associ- ated with poor overall performance as represented by the EF composite.

Nonexecutiveftlnction. Increased motor latency was marginally associated with higher phenylalanine lev- els at ages 9 to 13 and 14 to 15. Low memory spans were associated with high phenylalanine levels at ages 9 to 13, 14 to 15, and in the past 6 months. High phenylalanine levels at ages 9 to 13 also related to cur-

Table 4 Associations among Age, IQ, and Biochemical Indices

Age When Blood Levels of Phe and Tyr Were Recorded

0-2 Years 3-4 Years 5-8 Years 9-13 Years 14-15 Years Past 6 Months

Age Phelage .17 -.27 .45* .58** .64* .61* Tyrlage .51+ -.41+ -.88** -.25 -.44* -.31 Ratio/ age -.46 -.07 .76** .74** .75** .74**

IQ Phe/ IQ -.49* -.47* -.71** -.75** -.56** -.69* Tyr/IQ -.20 .38+ -.OO -.26 .02 .23 Ratio/ IQ -.39+ -.57* -.52* -.54* -.48* -.39+

Note: Age and IQ correlations are Pearson correlations. Bold-faced values indicate correlations that conform to the expectations that high phenylalanine (Phe), low tyrosine (Tyr), and/or a high Phe:Tyr ratio will be associated with lower 1~sand increased age. Note that values recorded during ages 0-2 exclude those used to diagnose phenylketonuria.

*p < .05; **p< .01; +p < .lo.

Table 5 Associations among Executive Function (EF) Task Performance, Phenylalanine (Phe), and Tyrosine (Tyr)

Age When Blood Phe and Tyr Levels Were Recorded

0-2 Years 3-4 Years 5-8 Years 9-13 Years 14-15 Years Past 6 Months

EF composite Phe/ EF Tyr/EF Ratio/ EF

Spatial working memory Strategy1 Phe -.I1 .38+ .23 .39+ .65* .71** Strategy /Tyr -.61+ .21 .16 .56* .22 -.02 Strategy/ ratio .89** .20 .20 .09 .43+ .56+ Errors / Phe -.44+ .26 .18 .30 .15 .13 Errors /Tyr -.81** -.20 .45+ .58* .22 .25 Errors / ratio .58+ .28 .08 .OO .38+ .17

Set shifting Stage/ Phe Stage/Tyr Stage/ ratio

Tower of London Perfect solutions /Phe .54* -.05 -.09 -.30 .04 -.30 Perfect solutions /Tyr .92** -.51* -.46+ -.I8 .04 .65* Perfect solutions / ratio -.58* .17 -.I3 -.25 -.39+ -.71*

Note: Values are partial correlation coefficients, covarying for age and IQ. Bold-faced values indicate correlations that conform to the
study's predictions that cognitive function will be impaired by high Phe, low Tyr, and/or a high Phe:Tyr ratio.
*p < .05; **p < .01; p < .lo; "#coefhcient cannot be determined because of lack of variance.


rent spatial recognition memory performance. Other- the minimum number of moves was associated with wise, correlations between phenylalanine levels and higher tyrosine levels at ages O to 2 years and in the measures of nonexecutive function were nonsigruficant. past 6 months. Counter to prediction, high tyrosine levels at ages 3 to 4 and 5 to 8 were associated with poorer Tower of London performance. The EF com-

Tyrosine and Cognitive Outcome posite was most related to tyrosine levels at ages O to

The range of most recently obtained tyrosine levels 2 years, with low tyrosine levels associated with varied from .50 to 1.7 mg / dl (M = 1.05,SD = .36). The poorer current performance. reference range for normal tyrosine levels in individ- Nonexecu tiuefunc tion. In terms of nonexecutive uals above the age of 12 years is from .6 to 2.4 mg/dl performance, longer motor response latencies were according to the University of Minnesota's Depart- associated with high tyrosine levels at ages 0 to 2 years. ment of Biomedical Genetics, where the assays were A relatively hgher spatial memory span was related to conducted. Tyrosine levels were correlated with all high tyrosine levels at ages 0 to 2 years. Finally, better cognitive task variables, partialing out the effects of pattern recognition memory was associated with both IQ and age on performance. These findings are higher tyrosine levels in the most recent period. represented in Tables 5 and 6.

Executiuefunction. Spatial working memory im-The Phe:Tyr Ratio and Cognitive Outcome

pairment, as measured by a high number of mnemonic errors and poor use of strategy, was associated As indicated in Table 2, phenylalanine and tyrosine with low tyrosine levels at ages O to 2 years. However, levels measured within the past 6 months were in- counter to prediction, higher tyrosine levels were as- versely correlated to a moderate degree, r(17) = -.50. sociated with poor performance on this task at ages 5 To address Diamond and colleagues' (1997) hypothesis to 8 (increased mnemonic errors) and 9 to 13(increased that the ratio between phenylalanine and tyrosine leads errors and high strategy scores). Stage reached on the to dopamine deficiency in the prefrontal cortex and set-shifting task was unrelated to tyrosine levels. A associated cognitive dysfunction, ratio scores were high number of Tower of London problems solved in correlated with all tasks, covarying once again for IQ

Luciana, Sullivan, and Nelson 1647

Table 6 Associations among Nonexecutive Function (NEF) Task Performance, Phenylalanine (Phe), and Tyrosine (Tyr)

Age When Phe and Tyr Levels Obtained

0-2 Years 3-4 Years 5-8 Years 9-13 Years 14-15 Years Past 6 Months

NEF composite Phe / NEF Tyr/NEF Ratio / NEF

Motor screening Accuracy / Phe Accuracy / Tyr Accuracy /ratio Latency /Phe Latency /Tyr Latency /ratio

Memory span Span/ Phe SpanITyr Span/ ratio

Recognition memory Pattems/ Phe Pattems/Tyr Patterns/ ratio Locations/ Phe Locations/ Tyr Locations/ ratio

Note: Values are partial correlation coefficients, covarying for age and IQ. Bold-faced values indicate correlations that conform to the
study's predictions that cognitive function will be impaired by high Phe, low Tyr, and/or a high Phe:Tyr ratio.
*p < .05; **p < .01; + p < .lo; nnncoefficient cannot be determined because of lack of variance.

and age. The findings, represented in Tables 5 and 6, suggest that the Phe:Tyr ratio is moderately to strongly associated with several aspects of cognitive function.

Executivefunction. Increased errors on the spatial working memory task were associated with increased Phe:Tyr ratios at ages 0 to 2 and 14 to 15. Poor use of strategy was also associated with a higher ratio score at these ages, as well as within the past 6 months. A decreased stage reached on the ID/ED set-shifting task was associated with higher Phe:Tyr ratios at ages 3 to 4,5 to 8,9 to 13, and 14 to 15. Similarly, a decrease in the number of Tower of London problems solved in the minimum number of moves was related to in- creased ratio scores at ages 0 to 2,14 to 15, and in the past 6 months. As expected from this pattern of find- ings within individual tasks, a high current EF com- posite score was associated with increased ratio scores at ages 0 to 2,5 to 8,9 to 13, and 14 to 15 years.

Nonexecutivefunction. With respect to nonexecutive functions, increased motor response latencies were associated with higher ratio scores at ages 14 to 15 and in the past 6 months. Higher ratio scores at ages 14 to 15 were also associated with poor pattern recog- nition memory performance. In the age range from 0 to 2 years, high ratio scores were associated with bet- ter current NEF performance. However, high ratio scores at ages 9 to 13 and 14 to 15 were associated with poorer NEF performance, as indicated by the composite index.

Implications of high versus low ratio scores. To better understand the implications of high Phe:Tyr ratio scores for current levels of cognitive function, the me- dian ratio score (represented by a value of 14.6) at ages 14 to 15 years was used to group PKU probands into those with high versus low scores. This age range was selected because all PKU probands had values available for this time period and because it was sig- nificantly associated with performance on a variety of tasks. The performance of these two groups was com- pared once again with that of the control samples. Findings indicated that PKU probands with high ratio scores performed significantly worse than the control sample in terms of the NEF composite, F(2,43) = 3.62, p < .05, and marginally worse in terms of the EF composite, F(2, 43) = 2.49, p < .lo. With respect to specific tasks, the high-ratio PKU sample performed worse than controls in terms of the maximum stage reached on the set-shifting task, F(2, 43) = 3.29, p < .05. All other specific task variables yielded nonsignif- icant differences between the high-ratio PKU, low- ratio PKU, and control samples.

A large number of correlations have been pre- sented in these analyses, and control for multiple cor- relations has not been stringently applied. In Tables 5 and 6, there are a total of 194 correlations reported. We utilized one-tailed tests of significance and have highlighted p values less than .lo. Hence, we would expect 10% of these correlations to be significant by chance. In Table 5, which represents 87 correlations between measures of executive function and three biochemical indices (phenylalanine, tyrosine, and the Phe:Tyr ratio), 31 correlations (or 36%) were signifi- cant in the predicted direction, a number that exceeds chance expectancies. Moreover, 48% of the correla- tions that were significant in the expected direction represent associations that involved the Phe:Tyr ratio. Table 6 presents 107 correlations between measures of nonexecutive function and the same biochemical in- dices. Here, 16 of 107 correlations (15%) were signifi- cant in what might loosely be described as the pre- dicted direction (e.g., poor cognitive performance associated with high phenylalanine and low tyrosine levels). Of these, 38% involved the ratio score. This was, admittedly, a crude analysis. Biochemical levels obtained at one age are highly associated with those obtained at past and subsequent ages, and the ratio score is related to each of the two indices that com- prised it. Moreover, the cognitive task variables, par- ticularly those derived from the same task, are signif- icantly intercorrelated. A larger study sample would permit the use of factor-analytic techniques to reduce the data into smaller components, but data reduction techniques were not appropriate with this sample. Thus, because of the number of comparisons reported, we stress that these findings await replication, although the observed pattern of effects was clearly in the predicted direction.


Most studies of cognitive function in early and con- tinuously treated PKU have included participants who are infants or preschoolers (Diamond et al., 1997). Long-term cognitive outcome is uncertain, par- ticularly in adolescents and young adults who have achieved normal IQ scores. The results of the current study suggest that with continuous dietary treatment, individuals with PKU who have achieved nor- mal IQ scores do not differ substantially from their peers or from an age-matched group of individuals with other chronic illnesses in terms of cognitive per- formance. The cognitive measures used in this study are particularly sensitive to frontostriatal dysfunc- tion, as evidenced by both clinical and functional neu- roimaging studies (Baker et al., 1996; Fray et al., 1996; Morris, Ahrned, Syed, & Toone, 1993; Owen, Doyon, Petrides, & Evans, 1996; Owen, Evans, & Petrides, 1996; Owen et al., 1992; Owen, Morris, Sahakian, Polkey, & Robbins, 1996; Owen, Roberts, Pokley, Sahakian, & Robbins, 1991; Veale, Sahakian, Owen, & Marks, 1996). Although relatively low scores on tasks that index frontal lobe function have been reported in younger samples of PKU-affected individuals, it cannot be generally concluded on the basis of the current data that adolescents and young adults with PKU will nec- essarily demonstrate a deficit level of performance on measures of executive function.

Despite what appears to be a relatively positive out- come for the PKU sample as a whole, dietary control of phenylalanine and tyrosine levels nevertheless ap- pears to contribute to individual differences in cogni- tive function. Diamond and colleagues (Diamond et al., 1997) have discussed the importance of the Phe:Tyr ratio as an index of the level of tyrosine that will be available to the brain for dopamine synthesis. A high ratio indicates that phenylalanine levels are in a range that would theoretically impede the transport of avail- able tyrosine across the blood-brain barrier, resulting in a decreased availability of tyrosine for dopamine synthesis. The findings of this study support the hy- pothesis that in individuals with PKU who may have decreased tyrosine availability, frontal lobe functions that are dopamine dependent may be compromised. High Phe:Tyr ratio scores in the past 6 months, at ages 14 to 15 and at ages 0 to 2 years were associated with relatively poor current spatial working memory per- formance (both in terms of number of errors and use of strategy) and a relatively poor ability to solve Tower of London problems with maximal efficiency. Relatively poor set-shifting performance was associated with high ratio scores in all age ranges between 3 and 15 years of age. Moreover, below-average scores on a composite index of EF performance were associated with high Phe:Tyr ratios at ages 0 to 2, 5 to 8, 9 to 13, and 14 to 15 years. However, counter to prediction, high ratio scores, particularly at ages 9 to 13 and 14 to 15, were also associated with relatively poor nonexec- utive function performance. Indeed, individuals with high ratio scores at ages 14 to 15 performed signifi- cantly worse than control participants in terms of the NEF composite. If these findings can be replicated, then the specificity of dopamine's role in mediating performance on tests of executive function, particu- larly in the context of what is hypothesized to be a life- time dopamine deficit, should be re-evaluated.

In addition, to better examine the significance of the Phe:Tyr ratio as a measure of dietary control in PKU, it would be important to examine cognitive per- formance in on- and off-diet groups, including groups of phenylketonurics who maintain dietary phenylala- nine restrictions but have discontinued tyrosine sup- plementation. Another interesting avenue of research would be to examine, through pharmacological chal- lenge studies, whether PKU probands with moder- ately high phenylalanine and low tyrosine levels are sensitized to the effects of catecholamine antagonists. Administration of dopamine antagonists to monkeys and humans results in measurable decrements in working memory abilities (Luciana & Collins, 1997; Sawaguchi & Goldman-Rakic, 1991). It would be pre- dicted that relative to normal controls, PKU probands who are hypothesized to be experiencing at least sub- tle reductions in dopamine would demonstrate an augmented response to drugs that further lower cate- cholamine levels in the prefrontal cortex and to ty- rosine depletion in particular. A study of this type is currently under way and might be useful in contrib- uting to the development of new and less invasive treatment strategies for the disorder.

Our findings are also important in contributing to the debate over whether lifetime dietary control of phenylalanine is necessary. Because participants in this study (as well as in other studies) who exemplify a cur- rent lack of dietary control also have histories of poor control, it is difficult to adequately determine the threshold above which a high Phe:Tyr ratio is harmful, and for how long a period of time. Indeed, whether transient increases in phenylalanine levels lead to mea- surable cognitive deficits in PKU probands with nor- mal IQ scores, who have otherwise been well-controlled in their lifetimes, is a matter of debate (Griffiths, Smith, & Harvie, 1997). Weglage et al. (1996) documented spe- cific deficits in selective and sustained attention that were significantly correlated with phenylalanine levels obtained on the day of testing. Additionally, there have been reports of improvement in neuropsychological functions in PKU probands who have not generally adhered to dietary phenylalanine restrictions when these individuals return to a low phenylalanine diet. Such improvements include faster reaction times and better sustained attention (Clarke, Gates, Hogan, Bar- ret, & MacDonald, 1987; Schmidt, Rupp, Burgard, Pietz, Weglage, & de Sonneville, 1994). Krause et al. (1985) demonstrated reversible changes in higher inte- grative functions in PKU probands tested under high versus low phenylalanine conditions, although these researchers did not consider the influence of past phe- nylalanine levels in the interpretation of their findings. Overall, our developmental data suggest that contin- uous dietary control of both phenylalanine and ty-

Luciana, Sullivan, and Nelson 1649

rosine throughout childhood and early adolescence is beneficial to cognitive function.

Several limitations of the current findings should be mentioned. First, PKU probands with more cogni- tive deficits may have been less likely to voluntarily participate in this study. Indeed, it appears that our PKU sample represents outcome at its very best in terms of the attainment of normal IQ scores and lack of deficits in executive function when compared with age-matched controls. In light of this aspect of sam- pling, the significance of our correlational analyses is perhaps all the more striking. Second, it is clearly a weakness of the current design that we did not mea- sure IQ in the comparison samples. The third issue to be raised concerns the nature of our EF versus NEF tasks and how task differences might have contrib- uted to associations between cognitive function and biochemical variables. It may be that the EF tasks were simply more difficult than the NEF tests. Cer- tainly the nature of many tests of executive function is that they require a high degree of "multitaslung," whereas tasks dependent on posterior brain regions often seem to be less complex in their demands (Dia- mond et al., 1997; Luciana et al., 1999). It should cer- tainly be the case that within healthy individuals, EF tasks will, by definition, be more challenging than NEF tasks, and available evidence suggests that chil- dren develop competencies on NEF prior to EF mea- sures (Luciana & Nelson, 1998). It is apparent from research using this battery of tasks that individuals with posterior brain lesions have relatively greater difficulty with the tasks we have classified as NEF tasks, while those with frontal lesions have greater difficulty with tasks we have classified as EF tasks (Owen, Doyon et al., 1996; Owen, Evans, & Petrides, 1996; Owen, Morris et al., 1996). Based on this litera- ture, if the PKU probands in this study had difficul- ties due to diffuse brain pathology (versus specific frontal disruptions), they should have demonstrated a pattern of performance that called into question their performance on both NEF and EF measures. Specifically, if processes of myelination or general structural organization were disrupted, as has been suggested by other researchers (Diamond et al., 1997), we would have expected group differences on mea- sures of reaction time, basic recognition memory for patterns and locations, memory span, and basic dis- crimination and reversal learning.

One final issue concerns the inherent confound be- tween executive competence and dietary compliance. That is, the cause-effect relation between high Phe:Tyr levels and executive performance is difficult to reli- ably determine. We have suggested that poor dietary control of phenylalanine levels leads to a state of com- promised ability on tests of executive function and that the mechanism underlying this effect is due to a cascade of events culminating in decreased dopamine function in the prefrontal cortex of affected individu- als. However, it may also be the case that deficient executive function occurs in PKU due to an as yet un- determined mechanism and that poor executive com- petence leads to a lack of compliance with dietary re- strictions, in turn leading to the observation of high phenylalanine levels and a high Phe:Tyr ratio. Since phenylalanine levels cannot be randomly controlled in human studies, one way to address this issue is to pursue animal models of the disorder, as has been done by Diamond and colleagues (Diamond, Ciar- amitaro, Donner, Djali, & Robinson, 1994). Regardless of the nature of the relation, it is certainly the case that individuals with poor executive control are theoreti- cally at higher risk for dietary noncompliance, and to the extent that dietary compliance is necessary for neurological and psychological health, these individ- uals are particularly vulnerable to additional cogni- tive impairment.

Despite these limitations, these findings suggest new avenues for research into the biochemical bases of PKU and perhaps other developmental disorders. It has yet to be explained why cognitive deficits remain in individuals who are early and continuously treated for the disorder. The hypothesis that a neurochemical (dopamine) dysfunction is present despite conven- tional treatment provides an explanation for these def- icits and should be further explored through empirical studies. Our data are consistent with past findings in that high phenylalanine levels are associated with rel- atively poor patterns of performance on cognitive tests in this small sample of PKU-affected individuals. We extend this literature by demonstrating that a high Phe:Tyr ratio is also consistently associated with poor performance on tests of executive function. Moreover, the level of cognitive proficiency demonstrated in young adulthood, and is associated with the ratio score at several points earlier in development. Whether di- etary restriction of phenylalanine should continue past childhood in individuals with PKU has been debated. These findings suggest that there are advantages to maintaining dietary control of both phenylalanine and tyrosine levels into early adulthood, and that dietary control should be evaluated, not only in reference to phenylalanine levels as is the convention, but also rel- ative to the magnitude of the Phe:Tyr ratio.


This work was partially supported by a research grant awarded to Jill Sullivan from the PKU Founda- tion at the Fairview-University Medical Center, and by a Junior Faculty Research Award to Monica Luci- ana from the MacArthur Foundation's Research Net- work on Psychopathology and Development (David Kupfer, MD, Chair). Thanks are extended to the fol- lowing individuals for their support in recruiting par- ticipants for this study: Drs. Robert Fisch and PiNian Chang of the PKU Clinic in the Department of Pediatrics at Fairview-University Medical Center, Dr. Warren Warwick of the Pediatric Pulmonary Clinic, Dr. An- toinette Moran of the Pediatric Endocrinology Clinic, and Dr. Paul Kubic of the Pulmonary Clinic at Chil- dren's Health Care in St. Paul, MN. The authors also thank the participants and their families for their in- terest in this research.


Corresponding author: Monica Luciana, Depart- ment of Psychology, 75 East River Road, University of Minnesota, Minneapolis, MN 55455; e-mail: .Jill Sullivan and Charles

A. Nelson are also at the University of Minnesota.


Ahlers, S. T., Salander, M. K., Shurtleff, D., & Thomas, J. R. (1992). Tyrosine pretreatment alleviates suppression of schedule-controlled responding produced by corticotro- pin releasing factor (CRF) in rats. Brain Research Bulletin, 29,567-571.

Baker, S. C., Rogers, R. D., Owen, A. M., Frith, C. D., Dolan,

R. J., Frackowiak, R. S. J., & Robbins, T. W. (1996). Neural systems engaged by planning: A PET study of the Tower of London task. Neuropsychologia, 34,515-526.

Benson, P. F., & Fensom, A. H. (1985). Genetic biochemical disorders. Oxford: Oxford University Press.

Bradberry, C. W., Karasic, D. H., Deutch, A. Y., & Roth, R. H. (1989). Regionally-specific alterations in mesotelence- phalic dopamine synthesis in diabetic rats. Association with precursor tyrosine. lournal of Neural Transmission, 78,221-229.

Brozoski, T. J., Brown, R. M., Rosvold, H. E., & Goldman, P. S. (1979). Cognitive deficit caused by regional depletion of doparnine in prefrontal cortex of rhesus monkey. Science, 205,929-932.

Butcher, J. N., Williams, C. L., Graham, J. R., Archer, R., Telle- gen, A., Ben-Porath, Y. S., & Kaemmer, B. (1992). Minnesota Multiphasic Personality Inventoy-Adolescent (MMPI-A): Manualfor administraion, scoring, and interpretation. Minneapolis: University of Minnesota Press.

Clarke, J. T. R., Gates, R. D., Hogan, S. E., Barret, M., &Mac- Donald, G. W. (1987). Neuropsychological studies on adolescents with phenylketonuria returned to phenylal- anine-restricted diets. American lournal of Mental Retarda- tion, 92, 255-262.

Coull, J. T., Middleton, H. C., Robbins, T. W., & Sahakian, B. J.

Luciana, Sullivan, and Nelson 1651

(1995). Constrasting effects of clonidine and diazepam on tests of working memory and planning. Psychopharmacology, 120,311-321.

Deijen, J. B., & Orlebeke, J. F. (1994). Effect of tyrosine on cognitive function and blood pressure under stress. Brain Research Bulletin, 33, 319-323.

Diamond, A. (1996). Evidence for the importance of dopamine for prefrontal cortex functions early in life. Philosophical Transactions of the Royal Society of London-Series B: Biological Sciences, 351, 1483-1493.

Diamond, A., Ciaramitaro, V., Donner, E., Djali, S., & Robinson, M. B. (1994). An animal model of early-treated PKU. Journal of Neuroscience, 14, 3072-3082.

Diamond, A., & Herzberg, C. (1996). Impaired sensitivity to contrast in children treated early and continuously for phenylketonuria. Brain, 119, 523-538.

Diamond, A., Prevor, M. B., Callender, G., & Druin, D. (1997). Prefrontal cortex cognitive deficits in children treated early and continuously for PKU. Monographs of the Society for Research in Child Development, 62(4, Serial No. 252).

Dias, R., Robbins, T. W., & Roberts, A. C. (1996). Dissociation in prefrontal cortex of attentional and affective shifts. Nature, 380, 69-72.

Downes, J. J., Roberts, A. C., Sahakian, B. J., Evenden, J. L., Morris, R. G., & Robbins, T.W. (1989). Impaired extra- dimensional sluft performance in medicated and unrned- icated Parkinson's disease: Evidence for a specific atten- tional dysfunction. Neuropsychologia, 27, 1329-1343.

First, M. B., Spitzer, R. L., Gibbon, M., & Williams, J. B. W. (1994). Structured Clinical lnterview for Axis-1 DSM-IV Disorders. New York: New York State Psychiatric In- stitue, Biometrics Research Department.

Fitts, W. H., &Warren, W. L. (1996). Tennessee Self-concept Scale (TSCS:2).Los Angeles: Western Psychological Services.

Fray, P. J., Robbins, T. W., & Sahakian, B. J. (1996). Neuro- psychiatric applications of CANTAB. Iizternational Journal of Geriatric Psychiatry, 11, 329-336.

Griffiths, P., Smith, C., & Harvie, A. (1997). Transitory hyper- phenylalanaemia in children with continuously treated phenylketonuria. American Journal of Mental Retardation, 102,27-36.

Giittler, F., & Lou, H. (1986). Dietary problems of phenyl- ketonuria: Effect on CNS transmitters and their possible role in behaviour and neuropsychological function. lournal of Inherited Metabolic Disease, 9,347-353.

Hughes, C., Russell, J., & Robbins, T. W. (1994). Evidence for executive dysfunction in autism. Neuropsychologia, 32, 477-492.

Krause, W., Halminski, M., McDonald, L., Dembure, P., Salvo, R., Freides, D., & Elsas, L. (1985). Biochemical and neuropsychological effects of elevated plasma phenyl- alanine in patients with treated phenylketonuria. Journal of Clinical Investigation, 75,40-48.

Lange, K. W., Robbins, T. W., Marsden, C. D., James, M., Owen, A. M., & Paul, G. M. (1992). L-dopa withdrawal in Parkinson's disease selectively impairs cognitive perfor- mance in tests of frontal lobe function. Psychopharmacology, 107,394-404.

Luciana, M., & Collins, P. F. (1997). Dopaminergic modulation of working memory for spatial but not object cues in nor- mal humans. Journal of Cognitive Neuroscience, 9,330-347.

Luciana, M., Lindeke, L., Georgieff, M., Mills, M., &Nelson,

C. A. (1999). Neurobehavioral evidence for working memory deficits in school-aged children with histories of prematurity. Developmental Medicine and Child Neurol- ogy, 41,521-533.

Luciana, M., & Nelson, C. A. (1998). The functional emer- gence of prefrontally-guided working memory systems in four-to-eight year-old children. Neuropsychologia, 36, 273-293.

Mazzocco, M. M., Nord, A. M., van Doorninck, W., Greene,

C. L., Kovar, C. G., & Pennington, B. F. (1994). Cognitive development among children with early-treated phe- nylketonuria. Developmental Neuropsychology, 10,133-151.

McCann, U. D., Penetar, D. M., Shaham, Y., Thorne, D. R., Sing, H. C., Thomas, M. L., Gillin, J. C., & Belenky, G. (1992). Effects of catecholamine depletion on alertness and mood in rested and sleep-deprived normal volun- teers. Neuropsychopharmacology, 8,345-356.

McCann, U. D., Thorne, D., Hall, M., Popp, K., Avery, W., Sing, H. C., Thomas, M., & Belenky, G. (1995). The effects of 1-dihydroxyphenylalanine on alertness and mood in a-methyl-para-tyrosine-treated healthy humans: Further evidence for the role of catecholamines in arousal and anxiety.Neuropsychopharmacology, 13,41-52.

McKean, C. M., & Petersen, N. A. (1970). Glutamine in the phenylketonuric central nervous system. New England Journal of Medicine, 283, 1364-1367.

Milner, B. (1964). Some effects of frontal lobectomy in man. In J. M. Warren & K. Akert (Eds.), Thefrontal granular cor- tex and behavior (pp. 313-334). New York: McGraw-Hill.

Milner, B. (1971). Interhemispheric differences and psy- chological processes. British Medical Bulletin, 27,272-277.

Morris, R. G., Ahmed, S., Syed, G. M., & Toone, B. K. (1993). Neural correlates of planning ability: Frontal lobe activa- tion during the Tower of London test. Neuropsychologia, 31,1367-1378.

Morris, R. G., Downes, J. J., Sahakian, B. J., Evenden, J. L., Heald, A., & Robbins, T. W. (1988). Planning and spatial working memory in Parkinson's disease. Journal of Neu- rology, Neurosurgery, and Psychiatry, 51,757-766.

Moy, S. S. (1995). Impaired acquisition and operant responding after neonatal dopamine depletion in rats. Pharmacology, Biochemistry and Behavior, 52,433-441.

Nasello, A. G., & Ramirez, 0. A. (1978). Brain catecholamine metabolism in offspring of amphetamine treated rats. Pharmacology, Biochemistry and Behavior, 9, 17-20.

Owasoyo, J., Neri, D. F., & Lamberth, J. G. (1992). Tyrosine and its potential use as a countermeasure to perfor- mance decrement in military sustained operations. Aviation, Space and Environmental Medicine, 63,364-369.

Owen, A. M., Downes, J. J., Sahakian, B. J., Polkey, C. E., & Robbins, T. W. (1990). Planning and spatial working memory deficits following frontal lobe lesions in man. Neuropsychologia, 28, 1021-1034.

Owen, A. M., Doyon, J., Petrides, M., & Evans, A. C. (1996). Planning and spatial working memory: A positron emis- sion tomography study in humans. European Journal of Neuroscience, 8, 353-364.

Owen, A. M., Evans, A. C., & Petrides, M. (1996). Evidence for a two-stage model of spatial working memory pro- cessing within the lateral frontal cortex: A positron emis- sion tomography study. Cerebral Cortex, 6,31-38.

Owen, A. M., James, M., Leigh, P. N., Summers, B. A., Mars- den, C. D., Quinn, N. P., Lange, K. W., & Robbins, T. W. (1992). Frontostriatal cognitive deficits at different stages of Parkinson's disease. Brain, 115,1727-1751.

Owen, A. M., Morris, R. G., Sahakian, B. J., Polkey, C. E., & Robbins, T. W. (1996). Double-dissociation of memory and executive functions in working memory tasks fol- lowing frontal-lobe excisions, temporal lobe excisions or amygdalo-hippocampectomy in man. Brain, 119, 1597- 1615.

Owen, A. M., Roberts, A. C., Pokley, C. E., Sahakian, B. J., & Robbins, T. W. (1991). Extra-dimensional versus intra- dimensional set-shifting performance following frontal lobe excisions, temporal lobe excisions, or amygdalo- hippocampectomy in man. Neuropsychologia, 29,993-1006.

Ozonoff, S. (2001). Advances in the cognitive neuroscience of autism. In C. A. Nelson & M. Luciana (Eds.), Handbook of developmental cognitive neuroscience. Cambridge, MA: MIT.

Palmour, R. M., Ervin, F. R., Baker, G. B., & Young, S. N. (1998). An amino acid mixture deficient in phenylala- nine and tyrosine reduces cerebrospinal fluid catechol- amine metabolites and alcohol consumption in vervet monkeys. Psychopharmacology, 136, 1-7.

Pennington, B. F., van Doorninck, W. J., McCabe, L. L., & McCabe, E. R. (1985). Neuropsychological deficits in early-treated phenylketonuric children. American Journal of Mental Deficiencies, 89, 467-474.

Petrides, M., & Milner, B. (1982). Deficits in subject-ordered tasks after frontal and temporal-lobe lesions in man. Neuropsychologia, 220, 249-262.

Rammsayer, T. H., & Vogel, W. H. (1995). AMPT partially reverses euphoric but not psychomotor effects of alcohol in humans. Psychobiology, 23,248-255.

Reich, W., Leacock, N., & Shanfeld, K. (1995). Diagnostic In- terview for Children and Adolescents-Revised-Adolescent Version (DICA-R-A). St. Louis, MO: University of Wash- ington.

Ris, M. D., Williams, S. E., Hunt, M. M., Berry, H. K., & Leslie,

N. (1994). Early-treated phenylketonuria: Adult neuro- psychologic outcome. Journal ofPediatrics, 124, 388-392.

Santana, C., Martin, L., & Rodrigues, D. (1994). Tyrosine in- gestion during rat pregnancy alters postnatal develop- ment of dopaminergic neurons in the offspring. Brain Research, 635, 96-102.

Sawaguchi, T., & Goldman-Rakic, P. S. (1991). Dl dopamine receptors in prefrontal cortex: Involvement in working memory. Science, 251,947-950.

Schmidt, E., Rupp, A., Burgard, P., Pietz, J., Weglage, J., & de Sonneville, L. (1994). Sustained attention in adult phenylketonuria: The influence of the concurrent phenylalanine-blood level. lournal of Clinical and Experi- mental Neuropsychology, 16,681-688.

Shallice, T. (1982). Specific impairments of planning. Philosophical Transactions of the Royal Society ofLondon Bulletin, 298,199-209.

Shukitt-Hale, B., Stillman, M., & Lieberman, H. R. (1996). Tyrosine administration prevents hypoxia-induced dec- rements in learning and memory. Physiology and Behavior, 59,867-871.

Shurtleff, D., Thomas, J. R., Ahlers, S. T., & Schrot, J. (1993). Tyrosine ameliorates a cold-induced delayed matching- to-sample performance decrement in rats. Psychopharmacology, 112,228-232.

Shurtleff, D., Thomas, J. R., & Schrot, J. (1994). Tyrosine re- verses a cold-induced working memory deficit in humans. Pharmacology, Biochemistry and Behavior, 47,935-941.

Smith, M. L., Klim, P., Mallozzi, E., & Hanley, W. B. (1996). A test of the frontal-specificity hypothesis in the cognitive performance of adults with phenylketonuria. Developmental Neuropsychology, 12,327-341.

Spreen, O., Tupper, D., & Risser, A. (1984). Human developmen- tal neurpsychology. New York: Oxford University Press.

SPSS, Inc. (1999). SPSS for Windows (Release 10.0.5) [Com- puter Software]. Chcago: Author.

Stemerdink, N. B. A., van der Molen, M. W., Kalverboer,

A. F., van der Meere, J., Huisman, J., de Jong, L. W., Slijper, F. M. E., ver Kerk, P. H., & van Spronsen, F. J. (1999). Prefrontal dysfunction in early and continuously treated phenylketonuria. Developmental Neuropsychology, 16,29-57.

Sullivan, J. (in press). Emotinal outcome of adolescents and young adults with early and consistently-treated PKU. Journal ofpediatric Psychology, 26.

Tam, S. Y., Elsworth, J. D., Bradberry, C. W., & Roth, R. H. (1990). Mesocortical dopamine neurons: High basal fir- ing frequency predicts tyrosine dependence of dopa- mine synthesis. Journal ofNeural Transmission, 81,97-110.

Veale, D. M., Sahakian, B. J., Owen, A. M., & Marks, I. M. (1996). Specific cognitive deficits in tests sensitive to frontal lobe dysfunction in obsessive compulsive disor- der. Psychological Medicine, 26,1261-1269.

Waisbren, S. E., Brown, M. J., de Sonneville, L. M. J., & Levy,

H. L. (1994). Review of neuropsychological functioning in treated phenylketonuria: An information processing approach. Acta Paediatrica, 407(Suppl.), 98-103.

Weglage, J., Pietsch, M., Fiinders, B., Koch, H. G., & Ullrich,

K. (1996). Deficits in selective and sustained attention processes in early treated children with phenylketo- nuria-results of impaired frontal lobe functions? European Journal of Pediatrics, 155,200-204.

Welsh, M. C., Pennington, B. F., Ozonoff, S., Rouse, B., & McCabe, E. R. B. (1990). Neuropsychology of early- treated phenylketonuria: Specific executive function deficits. Child Development, 61, 1697-1713.

Williamson, M. L., Koch, R., Azen, C., & Chang, C. (1981). Correlates of intelligence tests results in treated phe- nylketonuric children. Pediatrics, 68, 161-167.

Yokogoshi, H., & Nomura, M. (1991). Effect of amino acid supplementation to a low protein diet on brain neu- rotransmitters and memory-learning ability of rats. Physiology and Behavior, 50, 1227-1232.

  • Recommend Us