Combined cerebral blood flow effects of a cholinergic agonist (milameline) and a verbal recognition task in early Alzheimer’s disease


Julian Trollor, MD, FRANZCP, Neuropsychiatric Institute, The Prince of Wales Hospital, Randwick, Sydney, NSW 2031, Australia. Email:


Abstract  RU 35926/CI-979 (milameline) is a partial muscarinic agonist with promnestic effects in animal models. Preliminary animal studies suggest that this agent has the capacity to reverse cholinergic dysfunction and that it may impact on regional cerebral blood flow (rCBF). A total of 10 subjects with Alzheimer’s disease (AD) of mild severity underwent high resolution split-dose single photon emission computed tomography (SPECT) during performance of a verbal recognition and control task, both before and after 18 weeks treatment with melameline or placebo. SPECT images were coregistered with individual’s magnetic resonance imaging scans allowing extraction of rCBF values from multiple anatomical regions of interest (ROI). The effect of milameline was examined in eight individuals who were found after unblinding to be taking active drug. Effects of milameline were most apparent in the frontal regions, basal ganglia and thalamus. In the group as a whole, the greatest increase in rCBF due to milameline treatment was observed in the left globus pallidus. Response to milameline treatment was associated with increases in rCBF in the cingulate gyrus bilaterally, and less so for the left thalamus. Milameline-related increases in rCBF values were exaggerated by the verbal recognition task. Milameline has a demonstrable effect on cerebral blood flow in mild AD. Consistent with emerging animal data, the effects on rCBF appear most prominent in frontal and subcortical regions in AD subjects. The effects on rCBF appear to be augmented by the performance of a cognitively demanding task, raising the possibility that such tasks could assist in building an awareness of the functional neuropsychopharmacology of drugs designed for cognitive enhancement.


Despite the widespread use of cholinergic augmentation as a treatment for Alzheimer’s disease (AD), relatively little is known about the cerebral correlates of this strategy in humans. In particular, the effect of cholinergic agonists on cerebral metabolism and blood flow in humans has not been a major focus of study, despite the potential of these agents as enhancers of cognition. Only one study has evaluated the acute effects of nicotine administration on regional cerebral metabolic rate of glucose (rCMRgl).1 There have been no studies evaluating the long-term effects of administration of muscarinic cholinergic agonists on regional cerebral blood flow (rCBF).

RU 35926/CI-979 (milameline) is an orally active, non-selective partial agonist at the muscarinic cholinergic receptor.2 Animal evidence supporting the promnestic effects of milameline comes from its ability to reverse cognitive deficits in models of cholinergic dysfunction such as the scopolamine induced impairment in rats3 and adult rhesus monkeys4 and the basal forebrain lesion induced deficits in rats.2

Animal studies indicate that milameline has in vivo cholinomimetic activity in rodents and primates2,5 as well as modest ability to increase phosphoinositide (PI) accumulation in rat hippocampus.6 Despite its non-selectivity, milameline appears to possess in vivo central cholinomimetic activity at doses lower than those required to produce substantial cholinergic side-effects.2 Preliminary measurements using the hydrogen clearance technique from platinum electrodes embedded into the frontal cortex of rats suggest that milameline significantly increases rCBF in rats pretreated with methyl scopolamine.2 These data suggest that functional amelioration of deficits in cholinergic neurotransmission by administration of cholinergic agonists should improve rCBF parameters in AD subjects.

The present study provides preliminary evaluation of rCBF changes associated with milameline in AD subjects. The study was performed at a single site as an ‘add-on’ study, carried out during a multicentre 26-week double-blind, placebo-controlled Phase 3 clinical trial of milameline. The authors aimed to investigate with single photon emission computed tomography (SPECT) the effect of the cholinergic agonist milameline on rCBF in a group of subjects with AD. In particular, the authors wished to examine the interaction between drug-induced rCBF changes and rCBF changes induced by a verbal recognition memory task. The authors hypothesized that long-term treatment with milameline would result in a relative increase in rCBF in the prefrontal cortex of AD subjects, and that these increases would be most apparent in subjects who were ‘responders’ to milameline therapy. The authors also considered that drug-related increases in rCBF would be more robust during performance of a verbal recognition memory task compared with a control task.



Subjects were 10 right-handed individuals diagnosed with dementia according to Diagnostic and Statistical Manual, 4th Edition (DSM-IV) criteria who also met the National Institute of Neurological and Communicative Disorders and Stroke – Alzheimer’s Disease and Related Disorders Association (NINCDS-ADRDA) criteria7 for probable AD. Subjects had a minimum duration of symptoms of 1 year. Subjects had mild dementia according to DSM-IV criteria, with Mini-Mental State Examination (MMSE)8 score ≥18, self-sufficiency on activities of daily living, and Clinical Dementia Rating Scale9 sum-of-boxes less than 5. Subjects had reliable caregivers, good nutritional status, adequate visual and auditory acuity, a normal physical and neurological examination and a normal electrocardiogram and chest X-ray. All subjects had normal laboratory test results including full blood count, electrolyte levels, renal and liver function tests, thyroid function tests and serum B12 and folate levels. Subjects were excluded if there was evidence on magnetic resonance imaging of infarction or widespread small vessel disease or if significant risk factors for cerebrovascular disease (Hachinski ischaemia score of ≥4)10 were present. Subjects were also excluded if there was a history of neurological disorder other than AD, major psychiatric disorder including substance abuse or a history of head injury with loss of consciousness. Both the subject and caregiver gave written informed consent and the studies were carried out with the approval of the Ethics Committee of the South-eastern Sydney Area Health Service, the Prince of Wales/Prince Henry Hospital Radiation Safety Committee and the New South Wales Radiation Advisory Council.

Medication administration

The medication consisted of identical capsules containing either milameline 0.5 mg, 1 mg or placebo, administered four times per day. Subjects were randomized to receive either milameline 3 or 4 mg per day, or placebo. After initial titration over a 2-week period, the milameline dose remained constant for the duration of the study. Monitoring of compliance was performed by questioning the caregiver and counting capsules remaining in the blister pack at each visit.


All assessments and extraction of rCBF values were performed independent of the clinical trial in which the subjects were participating, and blind to the status of the patient in the clinical trial. Subjects were assessed in an identical manner prior to randomized assignment to either placebo or active groups and then again during the 18th week of milameline or placebo treatment. At each assessment, subjects underwent standardized clinical assessment using the authors’ Memory Clinic protocol,11 detailed neuropsychological evaluation as previously reported12 and split-dose activation SPECT scan. Magnetic resonance imaging (MRI) scanning was performed during the initial assessment only.

Single photon emission computed tomography and magnetic resonance imaging

SPECT images were obtained in an identical manner during screening and after 18 weeks of milameline or placebo treatment. The authors’ split-dose scans and verbal recognition memory activation task have been previously described in detail.12 In brief, scanning sessions were conducted at a fixed mid-morning time-slot using a Prism 3000XP triple-head gamma camera (Philips Medical Systems, Milpitas, CA, USA) with low energy ultra high resolution fan beam collimators using the following parameters: 128 × 128 matrix, 120 steps over 360 degrees, 40 s per step. System resolution (FWHM) was 6.7 mm at 10 cm. Subjects received an injection of 370 MBq of 99mTc-HMPAO (Ceretec, Amersham Health, UK) during each of two cognitive tasks designated as ‘control’ and ‘activation’. For the ‘control’ task, subjects viewed a word and a non-word (nonsense letter string) and were asked to indicate the target (real word) by button press. For the ‘activation’ task, subjects were shown word pairs in random order and asked to indicate which of the pair was from a prelearned 10-item word list. Degree of difficulty was standardized for each subject by prelearning of the target word list to a threshold of six correct words on delayed free recall. For each stimulus the parameters were stimulus duration 2.5 s, interstimulus interval 1.5 s. Task administration commenced 30 s prior to the radiotracer injection and continued for 3.5 min post injection.

All subjects underwent MRI scan of the brain on a 1.5 Tesla Siemens scanner (Siemens, Munich, Germany). A 3D volume acquisition protocol was used comprising a scout mid-sagittal cut for AC-PC plane alignment followed by coronal FSPGR sequence (TR 12.2 ms, TE 5.3 ms, FOV 25 cm, matrix 256 × 256, slice thickness 1.5 mm, no gap).

Image analysis

SPECT and MRI images were processed in a manner identical to that previously reported by the author group.13 Images were displayed using ANALYZE software (Biomedical Imaging Resource, Mayo Clinic, Rochester, MN, USA) on a Silicon Graphics, Indy R5000 workstation. Coregistration of MRI and SPECT images was achieved using the Automated Multimodality Image Registration (AMIR) software package developed by Ardekani et al.14 Standardized anatomical regions of interest (ROI) were traced on each subject’s MRI as previously published13 and transferred to the SPECT images to derive counts for each region, both prior to and during treatment with milameline. At each time point, rCBF data were obtained for subjects during performance of the control task and the activation task, using a subtraction process after correction for decay. Whole brain counts were obtained by application of a whole brain ‘mask’ to the SPECT scan after segmentation of the subject’s MRI scan. Extraction of rCBF data was performed prior to unblinding of the study.

Statistical analysis

Analysis was only undertaken on the eight subjects who were found, after unblinding at the completion of the study, to have been taking milameline. All analyses were performed on rCBF data normalized to whole brain counts using SPSS version 10.0.05 (SPSS Inc., Chicago, IL, USA). A Bonferroni correction was applied to correct for multiple comparisons. Neuropsychological data before and during drug treatment were compared. Individual test scores derived at each of two time points were expressed as an age-scaled z-score using Mayo Clinic norms15–21 and converted to neuropsychological indices representing seven cognitive domains as previously described.12

For descriptive purposes, the authors calculated the percentage change in rCBF over the predrug rCBF values using the formula: % rCBF change due to drug = (rCBFon drug − rCBFoff drug)/rCBFoff drug × 100. The effect of treatment with milameline on rCBF was assessed using a repeated measures anova for rCBF data normalized to whole brain counts, with task (activation or control) and drug (on and off milameline) as the two within-subject factors and response (responder or non-responder) as a between-subject effect. In order to minimise the risk of a type I error, functionally related ROI from both right and left hemisphere were considered together as a collection of dependent variables comprising six ‘composite’ regions (prefrontal, temporal, parietal, occipital, cingulate and subcortical). These composite regions were used in each preliminary analysis before examination of univariate statistics for each individual ROI.


After unblinding, rCBF data for eight subjects who were found to be taking milameline were entered into the analysis. Of these eight, one subject failed to perform the activation task during the second SPECT scan and another failed to perform the activation task during both SPECT scanning sessions. At enrollment, subjects were suffering AD of mild severity (mean MMSE, 21.6; SD = 4.4; range, 16–27). There was no significant effect of milameline treatment on MMSE. Although reaction time during the control and activation task was reduced on milameline, this reduction failed to reach statistical significance (see Table 1). Group comparison of neuropsychological indices prior to and after 18 weeks of drug treatment for the eight subjects on active treatment revealed no significant change in any neuropsychological index after correction for multiple comparisons. However, scrutiny of individual’s neuropsychological indicies revealed that some subjects had experienced substantial improvements in some areas. As a result, the authors chose to examine rCBF data based on whether subjects were categorized as ‘responders’ or ‘non-responders’. Response was arbitrarily defined as whether subjects showed improvement in three or more of the seven neuropsychological indices after receiving 18 weeks of drug treatment.

Table 1.  Effect of milameline on dementia rating scales and task performance
VariableMean (SD)Mean difference95% CI or differencetP
Before drugOn drugLowerUpper
  •  n = 8, d.f. = 7;

  •  n = 6, d.f. = 5. P = 2 tailed significance levels. CI, confidence interval; HDRS, score on Hamilton Depression Rating Scale, 21 item; MMSE, mini-mental state examination score; RT, reaction time; SD, standard deviation; Task Score, % of correct responses.

MMSE 21.6 (4.4) 21.3 (6.8) 0.37 −2.23 2.980.340.743
HDRS  3.6 (2.1)  2.6 (1.8) 1.00 −0.26 2.261.870.104
Control Task Score 90.2 (17.4) 87.4 (23.0) 2.75 −2.06 7.561.350.218
Activation Task Score 81.6 (14.2) 78.3 (18.8) 3.30−11.71 18.310.570.596
RT: control (ms)1065 (360) 948 (325)118−38.09273.341.790.117
RT: activation (ms)1694 (209)1448 (334)246−97.36589.711.840.125

Profile of responders and non-responders

According to the authors’ classification, four subjects showed response and four showed non-response to milameline treatment. Age, duration and severity of dementia, drug dose and scores on Hachinski ischaemia scale did not differ between responders and non-responders. Qualitatively, responders contrasted with non-responders in showing increases for neuropsychological indices of verbal memory, visual memory, construction and psychomotor speed (Table 2). Responders also appeared to show larger improvements in reaction time (mean improvement 414 ms for responders, 79 ms for non-responders). However, none of these differences were statistically significant.

Table 2.  Neuropsychological and task performance of responders versus non-responders
 Responders (n = 4)Non-responders (n = 4)
PremilamelineOn milamelinePremilamelineOn milameline
  •  n = 3.

  • MMSE, mini-mental state examination score; RT, reaction time; SD, standard deviation; Task score, % correct responses.

MMSE 21.254.11 21.256.90 22.005.35 21.257.72
Attention Index  7.422.92  7.042.47  9.173.438.924.10
Verbal Memory Index  3.000.36  3.631.01  4.792.154.132.01
Visual Memory Index  2.130.25  2.630.95
Construction Index  5.752.87  6.004.08  6.504.805.504.12
Language Index  6.255.32  6.004.24
Frontal Index  5.563.63  5.083.09  6.734.115.583.26
Psychomotor Speed  4.631.89  5.002.12  8.003.507.672.75
RT: control (ms) 978486 9294271152211 966250
Control Task Score 88.9 21.2 85.8 27.3 91.5 15.8 89.0 22.0
RT: activation (ms)182712014133571561.00203.471482384
Activation Task Score 83.8 16.7 84.1 21.8 79.5 14.7 72.6 17.6

Effect of milameline on regional cerebral blood flow

Several ROI showed increased rCBF due to milameline. These were apparent for the left globus pallidus, (F1,4 = 26.38, P = 0.004), right globus pallidus (F1,4 = 8.63, P = 0.03) and right anterior cingulate (F1,4 = 8.58, P = 0.03). However, these findings were significantly affected by the small sample size, and only the left globus pallidus remained significant at P < 0.05 after correction for multiple comparisons.

Interaction between drug response and changes in regional cerebral blood flow

An interactive effect was noted between rCBF values and response to treatment in specific regions (see Table 3 and Fig. 1). Response to drug treatment was associated with increased rCBF values in the cingulate as a whole (F1,4 = 553.66, P = 0.03). A similar observation was apparent for the left thalamus (F = 14.51, P = 0.02), which was no longer significant after correction for multiple comparisons. As can be seen in Fig. 1, this effect is predominantly observed during performance of the activation task as responders were observed to have a steep rise in rCBF values induced by the drug treatment, whilst non-responders demonstrated a steep decline in rCBF.

Table 3.  Regional cerebral blood flow profile of responders and non-responders during control and activation task
ROIrCBF during control task (n = 4)rCBF during activation task (n = 3)
Pre drugOn drug% difPre drugOn drug% difPre drugOn drug% difPre drugOn drug% dif
  1. % dif, percentage change in rCBF induced by chronic milameline administration; rCBF, regional cerebral blood flow; ROI, regions of interest.

Right hemisphere
 Medial prefrontal0.970.99 1.70.941.1219.21.001.00  0.0
 Dorsolateral prefrontal1.031.06−  4.7
 Orbital prefrontal1.020.99−3.70.961.00 3.60.960.98 2.40.980.96 −2.4
 Prefrontal pole0.920.96 4.40.990.92− −5.6
 Hippocampus0.920.92−0.50.950.93−2.60.810.9010.30.840.95 12.7
 Anterior temporal1.041.02− 12.2
 Posterior temporal1.041.08  8.5
 Anterior parietal1.011.00−  4.9
 Posterior parietal1.051.05−  4.7
 Anterior cingulate1.000.99−−0.50.971.1822.00.901.04 15.6
 Cingulate body1.071.08 13.4
 Posterior cingulate1.081.05−−− 11.8
 Occipital1.021.03− 1.2
 Putamen1.261.32 4.21.311.35 5.4
 Globus pallidus1.121.21− 14.7
 Thalamus1.231.22−−0.61.311.39 6.11.321.34 2.0
Left Hemisphere
 Medial prefrontal0.950.98−1.90.951.03 −0.3
 Dorsolateral prefrontal1.011.03 1.3
 Orbital prefrontal0.960.97−1.50.970.97 0.00.950.99 3.5
 Prefrontal pole0.890.96 7.00.990.92−− −9.9
 Hippocampus0.900.88−2.20.930.89−4.30.810.82 0.40.810.83 2.9
 Anterior temporal0.950.95 0.00.950.97 −7.6
 Posterior temporal0.980.97−1.00.981.02 4.1
 Anterior parietal1.001.00− 2.0
 Posterior parietal1.041.03− 2.6
 Anterior cingulate1.010.99− 4.70.921.0413.81.020.95 −6.2
 Cingulate body1.031.03− 2.0
 Posterior cingulate1.101.10− −7.5
 Occipital1.011.01 0.50.950.96− 6.4
 Putamen1.311.36 3.41.351.42 −1.4
 Globus pallidus1.161.11− 20.9
Figure 1.

Effect of response to milameline on regional cerebral blood flow in selected regions of interest during control and activation task (r = responder, nr = non-responder to milameline).

Regional cerebral blood flow changes induced by verbal recognition task

The repeated measures analysis demonstrated a significant effect of the verbal recognition task for the prefrontal (F1,4 = 17 929.56, P = 0.006) and temporal (F1,4 = 315.55, P = 0.04) regions as a whole, where verbal recognition was associated with increases in rCBF relative to that during the control task. In individual ROI analysis, verbal recognition was associated with increases in rCBF in the right prefrontal pole (F1,4 = 130.39, P = 0.0003), left prefrontal pole (F1,4 = 13.01, P = 0.02), left posterior temporal lobe (F1,4 = 17.70, P = 0.014), right posterior parietal region (F1,4 = 19.65, P = 0.011) and right thalamus (F1,4 = 9.00, P = 0.04). However, after correction for multiple comparisons, significance was only preserved in the right prefrontal (P < 0.05) and right posterior parietal (P < 0.05) regions.

Interaction of drug treatment, response and effects of verbal recognition

Qualitatively, the total number of ROI showing increases in CBF in response to milameline were similar for responders (18 of 34 ROI) and non-responders (20 of 34 ROI) during performance of the control task, but favored the responders somewhat (29 of 34 ROI vs 23 of 34 ROI) during performance of the activation task. The main contrast appeared to be in frontal ROI where for responders, mean rCBF values remained stable or improved in seven of eight frontal ROI during both the control and activation task (non-responders three of eight and four of eight, respectively). Mean percentage increases in rCBF values due to milameline were modest (<5%) in responders during the control task but were of greater magnitude during the activation task (up to 29.8%) especially in the anterior cingulate deep gray nuclei and some prefrontal regions.

Within the frontal ROI as a whole, there was a suggestion that response to milameline treatment to be associated with larger increases in rCBF as a result of verbal recognition (interaction effect activation × response [F1,4 = 174.39, P = 0.057]). Two further findings are of interest, but also failed to maintain significance after correction for multiple comparisons. For the right cingulate body there was an interaction between drug treatment and activation status (F1,4 = 17.44, P = 0.014). In this ROI, reduction in rCBF was seen in response to verbal recognition prior to drug treatment but rCBF increased at activation during drug treatment. An interaction effect between milameline treatment, activation and drug response was apparent for the left thalamus (F1,4 = 8.83, P = 0.041). For this ROI, non-responders showed a small increase in rCBF before milameline, but a substantial reduction in rCBF in response to the verbal recognition paradigm on milameline. Conversely, for the responders, a dramatic increase in activation effects was seen during treatment with milameline.


Effects of milameline on regional cerebral blood flow

This small study is the first to examine the effects of milameline on CBF in humans during the performance of a cognitive task. The results suggest that milameline treatment does have an augmentation effect on rCBF in AD subjects, albeit not a dramatic one. In the group as a whole, the largest increase in rCBF was observed in the left globus pallidus, with smaller effects on the right globus pallidus and the right anterior cingulate. Consistent with this was the observation that response to milameline was associated with increased rCBF in the cingulate as a whole, and to a lesser extent with some deep gray nuclei. Although not reaching statistical significance, other measures such as the reduction in reaction time seen in responders during performance of the activation task and the (non-significant) improvement in psychomotor speed on formal neuropsychological testing is consistent with mediation of drug effect via fronto-subcortical pathways.

There is a paucity of literature regarding the effects of milameline on rCBF in humans. In considering its likely effects, models of cholinergic deficit may be instructive, as the effects on cerebral parameters may be considered opposite to the projected effects of milameline. First, scopolamine administration has been show to consistently reduce rCBF in the frontal cortex of humans22–24 and is known to reduce rCMRgl in cortical regions, thalamus and basal ganglia25,26 in animal studies. Second, in an animal model of cholinergic deficit involving lesioning the nucleus basalis magnocellular, a reduction in rCMRgl is seen which is particularly apparent in frontal cortex.27,28 Taken together, these findings would predict that reversal of the cholinergic deficit with milameline in early AD would result in regionally specific increases in CBF in the prefrontal cortex and deep gray nuclei. This is consistent with the findings of the present study.

Examination of the cerebral effect of milameline in animals is also instructive. Acute administration of milameline to rats pretreated with methyl scopolamine increases frontal CBF.2 Consistent with this is the ability of milameline to reverse scopolamine-induced decrements in performance in rats3 and rhesus monkeys.4 Regional binding of milameline to muscarinic receptors using 11C-labeled milameline has shown peak uptake in the frontal lobes and striatum of rhesus monkey brains.29 Taken together, these animal studies are congruent with the authors’ finding that long-term administration of milameline induced rCBF increases in frontal-subcortical regions.

Effects of verbal recognition

Previous work has implicated the prefrontal cortex in both encoding and retrieval of episodic memory in younger subjects.30,31 The authors found bilateral activation in the prefrontal poles of their subjects, consistent with studies of older subjects in which bilateral prefrontal activation can be observed with retrieval from episodic memory.32–34 However, the finding was more robust in the right prefrontal regions, consistent with the retrieval asymmetry predicted by the Hemispheric Encoding and Retrieval Asymmetry model.35,36 Temporal and parietal lobe activation was seen in the present subjects in response to verbal recognition, despite the presumed pathological burden within these areas. Although such activation is not considered a characteristic response to verbal recognition, diffuse temporal activation, particularly left-sided, has been a feature of verbal recognition in healthy elderly subjects33 and in younger subjects.37 Activation of bi-parietal regions has been seen as a feature of auditory-verbal recognition31 and visual recognition,38 that is, it is seen as a feature of recognition per se rather than being modality specific. Further comparison with a healthy control group would be of assistance in determining how activation patterns may have been influenced by the pathological burden of AD.

Interaction between drug treatment and verbal recognition

The present results suggest an interaction between milameline treatment and task-induced rCBF changes. The small sample size limits conclusions that can be drawn from this data, as these findings failed to reach statistical significance. However, the findings are discussed here as they are tantalizing and are considered of potential relevance for future studies examining interactive effects between cognition enhancing drugs and cognitive paradigms. The present study not only suggests an association between response to milameline and increase in frontal blood flow, but also that rCBF alterations are more prominent in the frontal and subcortical regions during the high demand task of verbal recognition. These features were highlighted by the dramatic reversal of the polarity of rCBF in the right cingulate and left thalamus during the performance of the highly demanding verbal recognition task compared to the word/non-word task. Virtually nothing is known about the interactive effects of drug treatment and cognitive activation paradigms. There are no previous reports of drug effects studied during cognitive activation in AD subjects. The results raise the possibility that, in a larger sample, cognitive activation may assist in highlighting drug effects. The predominance of the effect in the anterior portions of the right hemisphere with the verbal recognition paradigm is intriguing. A permissive effect of the cholinergic agonist can be postulated in which the drug promoted an increase in rCBF in response to the recognition memory task. Two previous studies from the same investigators39,40 have examined the rCBF response to physostigmine challenge during performance of a working memory task for faces in young healthy subjects. In both these studies, enhanced performance was seen in association with reduction in rCBF in right prefrontal cortex, which is at odds with the current study. However, it is likely that observed interactions between cognitive tasks and cholinergic manipulation are likely to be heavily influenced by the task being employed, the age of the subjects, the presence of any disease process and the type of cholinergic manipulation employed.

Limitations of the current study

The major limitations of this study are the small sample size and lack of placebo control group. However, the present sample size is on par with many similar studies1,41–47 that have had a significant impact on understanding of the cerebral correlates of cholinergic manipulation. The subjects for the current study were enrolled from a single site involved in a multicentre international study. A larger sample may have been obtained by conducting this study across several sites, but their wide geographic rendered this approach impractical. It was fortuitous that after unblinding, 8 of the 10 subjects who consented to this ‘add-on’ study were taking active treatment, therefore, taking greatest advantage of the limited sample. The small sample made it impractical to examine the relationship between rCBF changes and changes in task performance with drug. This ability would have enhanced the understanding of the mediators of rCBF effects. Therefore, the above results must be interpreted with due caution. Other limitations of the current study include the time-consuming nature of individual ROI analysis for each subject, making automated analysis an attractive option for similar studies of drug effects in the future.


Using SPECT in subjects with mild AD, the authors have demonstrated rCBF alterations secondary to chronic administration of the cholinergic agonist milameline. Despite the small sample size the authors have demonstrated some specific drug effects, in particular, increases in rCBF in frontal and subcortical regions. The authors have proposed a dynamic interactive effect between their cognitive task and the milameline, such that CBF changes associated with chronic administration of the drug are magnified by cognitive activation. The authors conclude that the study of the rCBF effects of both acute and chronic administration of cognitive enhancing drugs may be aided by concurrent performance of a cognitive activation task. Future studies should consider the additional utility of activation procedures and their role in understanding the functional neuropsychopharmacology of medications designed for cognitive enhancement.


This study was supported by Roussel-Uclaf, Romainville, France, who partially funded the cost of the scans.