Frontal subregions mediating Elevator Counting task performance.

Deficits in sustained attention may lead to action slips in everyday life as irrelevant action sequences are inappropriately triggered internally or by the environment. While deficits in sustained attention have been associated with damage to the frontal lobes of the brain, little is known about the role of the frontal lobes in the Elevator Counting subtest of the Test of Everyday Attention. In the current study, 55 frontal patients subdivided into medial, orbital and lateral subgroups, 18 patients with posterior lesions and 82 healthy controls performed the Elevator Counting task. The results revealed that patients with medial and left lateral prefrontal lesions were significantly impaired on the task compared to healthy controls. Research suggests that patients with medial lesions are susceptible to competition from task irrelevant schema; whereas the left lateral group in the current study may fail to keep track of the tones already presented.

Sustained attention is the ability to maintain conscious processing of a series of repetitive events presented at a moderately slow pace over a prolonged period of time. Without the ability to sustain attention, the presentation of such non-stimulating, repetitive events would lead to habituation and distraction to other stimuli (Robertson, Manly, Andrade, Baddeley, & Yiend, 1997). In everyday life, slips in sustained attention in healthy individuals might lead to putting salt instead of sugar on your cereal or adding milk to a coffee that was requested black (Reason, 1977). Such slips are thought to occur when attention to the task in hand diminishes due to monotony, anxiety or attending to more than one task at a time.

A number of tests have been devised to assess sustained attention in both healthy individuals and neurological patients. Examples include the Sustained Attention to Response Test (Robertson et al., 1997) and the Elevator Counting subtest from the Test of Everyday Attention (TEA; Robertson, Ward, Ridgeway, & Nimmo-Smith, 1994). In such tasks, individuals are required to either respond to all the digits in a sequence except for the digit 3 or count the number of events which are presented at a slow rate so that attention deteriorates. Typically, healthy individuals’ produce longer reaction times or the number of events missed increases as the time on task increases (Jerison & Pickett, 1963; Mackworth, 1950).

Vigilance network theories implicate the right fronto-parietal-thalamic network in the maintenance of sustained attention (Parasuraman, Warm, & See, 1998; Stuss & Benson, 1984). Indeed, patients with frontal lobe lesions are impaired on tasks assessing sustained attention (Molenberghs, Gillebert, Schoofs, Dupont, Peeters, & Vandenberghe, 2009; Shallice, Stuss, Alexander, Picton, & Derkzen, 2008; Wilkins, Shallice, & McCarthy, 1987). Moreover, neuroimaging studies have shown activation in the frontal lobes while healthy volunteers perform sustained attention tasks (Lawrence, Ross, Hoffmann, Garavan, & Stein, 2003; Pardo, Fox, & Raichle, 1991). In particular, the frontal lobes, including the anterior cingulate, are thought to be important for detecting targets and executive control (Posner & DiGirolamo, 1998; Posner & Petersen, 1990), the right frontal and bilateral parietal areas are associated with sustaining attention over time (Pardo et al., 1991; Posner & Petersen, 1990), and the thalamus, midbrain, and reticular system are associated with cortical arousal (e.g. Paus et al., 1997).

The TEA was devised as an ecologically valid test of different aspects of attention (Robertson et al., 1994). The Elevator Counting subtest of the TEA assesses the ability to sustain attention by presenting a long series of tones at a slow rate where the most favorable strategy for successful performance is to count the tones rather than estimate their number (Cordes, Gelman, Gallistel, & Whalen, 2001). Studies have demonstrated impairments in frontal patients on similar sustained counting tasks. For example, Wilkins et al. (1987) found that frontal patients failed to accurately count the number of events presented when a slow but not a fast presentation rate was used. However, to our knowledge, studies have not investigated frontal patients’ performance specifically on the Elevator Counting subtest.

Stuss, Shallice and colleagues propose that the distinct processes underlying sustained attention are associated with different frontal regions (e.g. Stuss, Shallice, Alexander, & Picton, 1995). Initiating and maintaining the intention to respond are thought to be associated with the superior medial frontal region (e.g. Alexander, Stuss, Shallice, Picton, & Gillingham, 2005). In contrast, monitoring the consequences of one's actions or comparing the status of the environment against task goals is associated with the right lateral prefrontal cortex (Picton, Stuss, Shallice, Alexander, & Gillingham, 2006). In a similar sustained counting task, Shallice et al. (2008) found that it was the superior medial frontal group who was significantly impaired compared to controls. The aim of the current study was to compare sustained attention abilities in a group of patients with focal frontal lesions and define which frontal subregions are critically implicated in performance on the Elevator Counting subtest. Patients were subdivided into medial, orbital and lateral subgroups according to the frontal region with greatest damage and performance on the Elevator Counting subtest was compared. It was envisaged that patients with lesions in the medial and the right lateral prefrontal region might be impaired on the task but not the other frontal groups.

Methods

Participants

Fifty five patients with frontal lobe lesions were recruited for the study from the National Hospital for Neurology and Neurosurgery. Patients were included if their lesion was confined to the frontal lobes, English was their first language, they did not have a severe disability which would prevent them from reliably responding in the interview such as severe dysphasia or neglect syndrome, and they did not have any other neurological or psychiatric disorders. Patients were identified by a neurologist on the basis of clinical MRI scans (or CT scans where MRI was unavailable) and the scans were coded for the presence (1) or absence (0) of a lesion in the three main anatomically defined subregions of the frontal lobes within each hemisphere (lateral, medial and orbital). Frontal lesions were localised by operation site in the case of surgical patients or by gross lesion characterisation in the nonsurgical patients. An area was only coded as damaged if at least 25% of that area was affected. The frontal patients were allocated either to the left lateral (n = 10), right lateral (n = 9), medial (n = 21) or orbital (n = 15) subgroups according to the region of greatest damage. The aetiologies within the subgroups were as follows – left lateral: glioma = 5; meningioma = 2; subarchnoid haemorrhage (SAH) = 2; metastasis = 1; right lateral: glioma = 4; meningioma = 4; SAH = 1; medial: glioma = 12; meningioma = 3; SAH = 4; traumatic brain injury (TBI) = 1; metastasis = 1; orbital: glioma = 1; meningioma = 2; abscess = 1; SAH = 11.

Eighteen patients with posterior lesions were also recruited for the study according to the same inclusion criteria except their lesions did not involve the frontal region. The patients’ aetiologies were: glioma = 9; meningioma = 7; temporal lobectomy = 1 with 55% with temporal lobe lesions, 17% with temporoparietal lesions, 17% with parieto–occipital lesions and 11% with parietal lobe lesions. The patient groups did not significantly differ in the mean time since surgery/onset (p = .21).

The performance of the patient groups was compared with 82 healthy controls. None of the controls had any previous history of head injury or stroke, major neurological or psychiatric illness, or alcohol abuse. A one-way analysis of variance (ANOVA) demonstrated that the patient groups and the control volunteers did not differ significantly in terms of age (p = .63) or years of education (p = .92). All participants were native English speakers. Consent was obtained according to the Declaration of Helsinki and the study was approved by the National Hospital for Neurology and Neurosurgery & Institute of Neurology Joint Research Ethics Committee. Table 1 shows the demographic and clinical data for the patients and controls.

Table 1

Demographic and clinical data for the patient groups and healthy controls.

Group	Age	Years of education	Gender, M/F	Handedness, right/left	Time since surgery/onset (months)	NART verbal IQ	Raven's APM (/12)	GNT (/30)	Verbal fluency, total words	Stroop CW time (s)	Frag. Letters (/20)	Elevator Counting (/7)
Frontal	46.90 (14.12)	13.21 (3.23)	33/22	48/7	11.10 (22.17)	104.91 (15.07)	7.71 (2.85)	19.49 (5.78)	30.41 (15.24)	177.18 (103.10)	19.42 (0.99)	6.33 (1.23)
Left lateral	49.80 (13.73)	12.90 (3.04)	5/5	9/1	5.63 (11.31)	108.50 (18.30)	7.50 (2.72)	20.90 (4.18)	23.40 (16.41)	261.83 (213.38)	19.50 (0.97)	5.80 (1.14)
Right lateral	47.67 (12.12)	13.67 (3.50)	4/5	7/2	6.22 (12.53)	113.22 (11.30)	8.67 (2.18)	22.33 (3.87)	38.78 (13.83)	170.70 (70.28)	19.44 (0.53)	6.67 (0.71)
Medial	43.57 (12.66)	13.43 (3.25)	11/10	18/3	18.00 (30.56)	103.62 (15.35)	7.90 (3.21)	19.43 (6.28)	31.05 (15.00)	141.33 (43.72)	19.19 (1.37)	6.10 (1.64)
Orbital	46.73 (17.95)	13.47 (3.44)	13/2	14/1	8.00 (16.73)	102.47 (12.45)	7.93 (1.91)	16.93 (6.24)	32.33 (13.32)	169.45 (64.44)	19.67 (0.49)	6.80 (0.56)
Posterior	51.22 (14.92)	12.61 (3.18)	8/10	15/3	10.44 (20.50)	104.61 (13.47)	6.76 (2.75)	18.50 (7.47)	33.28 (12.87)	165.98 (98.15)	19.65 (0.49)	6.56 (1.10)
Healthy controls	49.40 (15.42)	13.49 (3.09)	41/41	72/10	N/A	112.26 (9.61)	8.80 (2.13)	22.65 (4.21)	47.79 (13.05)	129.21 (30.94)	19.35 (0.79)	6.89 (0.35)

M/F, male/female; NART, National Adult Reading Test; APM, Advanced Progressive Matrices; GNT, Graded Naming Test; CW, Colour-Word; Frag. Letters, Fragmented Letters.

Procedure

Background neuropsychological measures. All participants performed the National Adult Reading Test-Revised (NART) to estimate premorbid levels of functioning (Nelson & Willison, 1991) and Raven's Advanced Progressive Matrices (APM) to assess nonverbal abstract reasoning (Raven, Raven, & Court, 1998). Naming abilities were assessed using the Graded Naming Test (GNT; McKenna & Warrington, 1983) and perceptual abilities using the Fragmented Letters subtest from the Visual Object and Space Perception Battery (VOSP; Warrington & James, 1991). Executive abilities were assessed using Controlled Oral Word Association (COWA; letters F, A and S; Spreen & Strauss, 1998) and the Stroop Test Colour-Word score (Trenarry, Crosson, Deboe, & Leber, 1989).

Elevator Counting subtest from the Test of Everyday Attention (TEA). The Elevator Counting subtest version C was administered according to the procedure in the manual (Robertson et al., 1994). Participants were asked to imagine that they were in a lift where the door indicator was not working. For each trial, participants had to imagine that they had entered the lift on the ground floor and determine which floor they have arrived at by counting the series of tones presented on a tape. There were 7 trials in total, with strings between 5 and 14 tones in length and random intervals of between 2 and 5 s between each tone (M = 3.45, SD = 1.00). Fixed sequences of tone length and intervals between tones were used.

Results

Background neuropsychological measures. A Kruskal–Wallis Test revealed that there was a significant main effect of group on NART performance [H(5) = 15.43; p < .01]. Mann–Whitney's U-Tests revealed that the orbital [U = 310.50; z = −3.04; p < .005], medial [U = 558.00; z = −2.48; p < .05] and posterior [U = 492.50; z = −2.21; p < .05] groups read significantly fewer words correctly than the healthy controls. However, the two lateral groups did not significantly differ from the controls (p > .85). On Raven's APM, there was also a significant main effect of group [H(5) = 11.16; p < .05] where the posterior patients had significantly poorer abstract reasoning compared to the control group [U = 382.50; z = −2.95; p < .005] but the frontal groups did not (p > .06). There was a significant main effect of group on the GNT [H(5) = 16.39; p < .01] where the orbital [U = 292.50; z = −3.23; p < .005] and medial frontal [U = 599.50; z = −2.15; p < .05] groups and the posterior group [U = 482.50; z = −2.30; p < .05] were significantly less accurate at naming compared to the controls while the other groups were not (p > .19). On the executive measures, there was a significant main effect of group on the COWA [F(5, 149) = 12.03; p < .0001]. All the frontal subgroups (all p < .0001) except the right laterals (p = .14), as well as the posterior patients (p < .0001), produced significantly fewer words on verbal fluency compared to controls. There was no effect of group on Stroop performance (p = .09) or the Fragmented Letters test (p = .56).

Elevator Counting subtest. The means for the patient groups and healthy controls performing the Elevator Counting subtest from the TEA are in Table 1. Kruskal–Wallis Test was conducted [H(5) = 26.98; p < .0001]. Subsequent pairwise analyses using the Mann–Whitney U-Test revealed that both the left lateral [U = 153.50; z = −5.01; p < .0001] and medial [U = 607.50; z = −3.30; p < .01] groups performed significantly more poorly than the healthy controls. In contrast, the right lateral, orbital and posterior groups did not differ significantly from controls (p > .23).

As some lesions involved more than one frontal area, the Elevator Counting task score was entered as a dependent variable into a linear model similar to that described by Müller-Plath, Ott, and Pollmann (2010). The dichotomous variables indicating the presence or absence of a lesion in the left lateral, right lateral, medial, orbital and posterior regions were entered as predictors. The best model of the regions most predictive of impairments in performance (F = 12.204, p < .001) included the left lateral (β = −0.28, p < .001) and medial frontal regions (β = −0.28, p < .001) and accounted for 0.14 of the variance. All other predictors were not significant.

Correlations between Elevator Counting subtest performance and other neuropsychological measures. Spearman correlation coefficients were calculated for the left lateral and medial groups to examine the relationship between the patients’ performance on the Elevator Counting subtest and the background measures. For the left laterals, there were significant positive correlations between performance on the Elevator Counting subtest and both GNT performance [r = .69, p < .05] and verbal fluency [r = .65, p < .05]. These correlations meant that the higher the score on the Elevator Counting subtest, the better the performance on the GNT and verbal fluency. No other correlations were significant.

Discussion

Previous studies have demonstrated that frontal patients are impaired in their ability to sustain attention when repetitive events are presented at a slow pace over a prolonged period of time (Wilkins et al., 1987). In a recent study, it was found that patients with superior medial frontal lesions were significantly impaired compared to healthy controls when both fast (3 tones/s) and slow (1 tone/3 s) presentation rates were used. Patients with right lateral lesions were impaired only when fast presentation rates were adopted while patients with orbital and left lateral lesions were not impaired in either case (Shallice et al., 2008). In the current study, it was the left lateral and medial frontal groups who were significantly impaired on the Elevator Counting subtest while the right lateral, orbital and posterior groups did not significantly differ from healthy controls.

In the current study, the mean interval between tones was 3.45 s which is similar to the slower presentation rate used by Shallice et al. (2008). Indeed, in line with Shallice et al. (2008) findings, our patients with medial lesions were impaired in their ability to sustain attention when repetitive events were presented at a slow pace. The medial group made 13% errors in comparison to 1.6% errors made by the controls. In contrast, our right lateral patients were not impaired on the Elevator Counting task. Again, this replicates Shallice et al. (2008) result that patients with right frontal lesions do not appear to be impaired on tasks of sustained attention when a slow rate of presentation is used.

Using Norman and Shallice's (1980) Supervisory Attention System approach, Stuss et al. (1995) have argued that when a task involves repetitive events presented at a slow pace, the control schema mediating the appropriate behaviour for that event becomes increasingly susceptible to competition from other task irrelevant schema triggered internally or by the environment. To prevent this from happening, the medial prefrontal cortex is held to contain systems which energise active supervisory processes to bias the competition in favour of the task relevant schema.

Unlike the study of Shallice et al. (2008), which involved only chronic patients, our current findings also revealed that the left lateral group was impaired on a task devised to assess sustained attention. The involvement of the left lateral prefrontal region may be due to the task placing demands on executive processing or cognitive control (Sylvester et al., 2003), although the left laterals were not impaired on the other executive measure administered. It may be that the left lateral group fail to use effectively or to initiate at the appropriate time counting strategies when performing the task. Indeed, the left lateral group's poor performance on verbal fluency and naming would support these speculations rather than a sustained attention deficit per se. In addition it could be that the left lateral group have difficulty updating and maintaining the contents of their working memory and so are unable to keep track of the tones already presented (Fletcher & Henson, 2001).

The current findings support the notion that the medial prefrontal cortex is important when a series of repetitive events is presented at a moderately slow pace over a prolonged period of time (Posner & Petersen, 1990; Shallice et al., 2008). It has been suggested that patients with lesions in the medial prefrontal cortex are not able to initiate processes necessary to cope with low rates of input (Paus et al., 1997; Stuss et al., 1995). However, both the current study and that of Shallice et al. (2008) failed to provide evidence to support Posner and Petersen's (1990) claim that the right lateral prefrontal cortex also plays a role in sustaining attention in such situations (but see Fan, McCandliss, Fossella, Flombaum, & Posner, 2005). Instead, impairments were found with left lateral prefrontal lesions. While this may be due to difficulty initiating or updating the counting of tones presented, future work should investigate this deficit in vigilance further.