A jury in the Superior Court found the defendant, Thomas E. Toolan, III, guilty of murder in the first degree of Elizabeth Lochtefeld.1 The murder conviction was based on *453theories of deliberate premeditation and extreme atrocity or cruelty. Before us is the defendant’s appeal from his convictions and the denial of his amended motion for a new trial. The defendant claims a number of separate errors related to the selection of the trial jury, the conduct of the trial, and the subsequent denial of his motion. We conclude that flaws in the jury selection process require the reversal of his convictions. We also consider briefly one other claim raised by the defendant that may arise at retrial.
1. Background. The jury could have found the following facts. In the fall of 2004, the victim, aged forty-four years, was living on Nantucket, as were her parents and brother. The defendant lived in New York City. The victim and defendant were introduced by a mutual friend on Nantucket on September 4, 2004. They became romantically involved, but in late October, during a visit to New York City, the victim ended the relationship because of the defendant’s heavy drinking. Leaving the defendant behind in New York, the victim returned to Nantucket on Saturday, October 23, 2004.
The next day, a security officer at LaGuardia International Airport stopped the defendant from carrying a knife through passenger screening on his way to catch a flight to Nantucket. The knife was seized and the defendant was given a criminal summons for carrying a knife over four inches long. The officer suspected that the defendant had “had a few drinks” and noticed that he smelled of alcohol. The defendant missed the flight and spent the night in the airport.
The following morning, Monday, October 25, the defendant flew to Nantucket. He arrived in Nantucket around 10:45 a.m., rented a car, and bought two knives. He drove to Hawthorne Lane, where the victim rented a cottage, and asked the victim’s neighbor and landlady, who lived next to the cottage, if anyone was at home there. Some time after noon, the neighbor grew concerned when she noticed that one window shade had been lowered and the victim had not yet left the house for a 1 p.m. appointment. The neighbor telephoned the victim’s brother, who in turn contacted the police. Meanwhile, the defendant took a 1 p.m. flight to Hyannis, where, at 1:21 p.m., he rented a second car.
*454In response to a 1:15 p.m. dispatch, police officers arrived at the cottage to conduct a well-being check and found the victim lying motionless on the living room floor. Shortly thereafter, a medical examiner arrived and pronounced the victim dead. The cause of death was multiple stab wounds to the victim’s chest and back.
Police quickly followed up on leads, including descriptions from neighbors of a rental car parked near the victim’s cottage, and of a young man in a fedora style hat. At the Nantucket airport at 3 p.m., police discovered a rental car matching the neighbors’ description. At 3:50 p.m., a Rhode Island State trooper on routine patrol on Route 95 received a broadcast to be on the lookout for a man in a floppy hat in a Chevrolet Impala automobile with a particular registration number. The trooper observed the car and the defendant in it. The trooper contacted other troopers and followed the defendant until the defendant was stopped at a roadblock set up for the purpose by police officers in Hopkinton, Rhode Island. The police stopped the defendant because he was a suspect in a homicide, but did not inform him at the time of the motor vehicle stop of the reason. The defendant was arrested, and an officer observed that he was somewhat lethargic, was unable to respond to simple commands, and smelled of alcohol.
The police transported the defendant to the Hope Valley barracks in Richmond, Rhode Island, and informed him he was suspected of driving a vehicle while under the influence of intoxicating liquor or drugs. Shortly before 7 p.m., after a breathalyzer test showed his blood alcohol level was over the legal limit for driving in Rhode Island, the defendant was charged with driving while under the influence of alcohol. That evening, members of the Massachusetts State police interviewed the defendant at the Rhode Island police barracks. He repeatedly asked the officers, “What’s this about?” and claimed not to have seen the victim since Friday.
Police found blood on the rental car left at the Nantucket airport. They matched blood from paper towels found at the airport and from the defendant’s clothing to the victim’s deoxyribonucleic acid (DNA). Subsequently, the defendant was charged with murder and assault and battery by means of a dangerous weapon.
*455The defendant’s trial began in the Superior Court for Nantucket County on June 4, 2007. The defense argued that the defendant was not criminally responsible for the victim’s murder due to brain damage and mental illness caused by the defendant’s addiction to alcohol, exacerbated by alcohol and prescription drugs ingested on the day of the victim’s murder. On June 21, 2007, the jury found the defendant guilty of both crimes.
2. Impartial jury. Prior to trial, in September, 2006, the defendant filed a motion for change of venue based on extensive pretrial publicity surrounding his case. He renewed the motion in court on the first day of jury empanelment. At the conclusion of jury empanelment, the judge denied the defendant’s motion. On appeal, the defendant argues that the trial judge’s refusal to grant a change of venue, to ask required questions of potential jurors, and to excuse unqualified jurors violated his right to a fair trial by an impartial tribunal, warranting a new trial under G. L. c. 278, § 33E. To provide context necessary to assess these arguments, we set out additional facts from the record.
a. Community context and media attention. Nantucket is a small island community: a 2006 town census counted 10,783 residents.2 Although the victim had only moved to Nantucket in the spring of 2004, she had spent many summers there since childhood, and several family members resided on the island. Her brother lived on Nantucket with his wife, an artist who showed her work locally, and their young children. Her parents also lived on the island; her father, also an artist, owned a gallery in town. The family was well known in the community.
The victim’s murder on October 25, 2004, was the first murder on Nantucket in over twenty years, and it attracted enormous attention. In the days and weeks following the murder, local and national media outlets published numerous articles highlighting the victim and defendant’s Nantucket-Manhattan “romance” and their respective personal histories. On November 15, 2004, People Magazine, a publication with a national circulation, ran a cover story called “Looking for Love, Finding Tragedy,” in which a former girl friend described the defendant as “a Jekyll *456and Hyde.” The article also described how, when the victim tried to end her relationship with the defendant a few days before the murder, he “became violent and kept her from leaving his apartment. She apparently managed to slip out at 4 a.m. the next day, [October] 23, when he was asleep, and fled back to Nantucket.” The alleged false imprisonment, which was never introduced at trial, appeared in one of Nantucket’s local newspapers, The Inquirer and Mirror, on November 6, 2004.
In 2006, a “true crime” book on the victim, the defendant, and the murder entitled, “Safe Harbor: A Murder in Nantucket,” was sold in Nantucket bookstores. On the first page of the introduction, the author wrote: “Underneath a self-admitted alcohol problem lurked a lethal hatred of women. For much of his adult life, it seemed, [the defendant] was a homicide waiting to happen.” Within the first twenty pages, the book stated that during their relationship, the defendant reacted angrily to the victim discovering a pistol in one of his drawers and, in a separate incident, put the victim “in a headlock” while drunk. The author also described the alleged incident in which the defendant held the victim captive in his apartment, including details that had not appeared in earlier magazine stories.
In November, 2006, approximately six months before the start of trial, articles published in the Inquirer and Mirror chronicled the filing of and hearing on the defendant’s motions to suppress evidence, and recounted in detail the testimony introduced at the hearing. In January, 2007, Nantucket’s other local newspaper, The Nantucket Independent, reported the judge had denied all six pretrial suppression motions because he found the defendant’s statements to the police while in custody in Rhode Island were voluntary, and the seized evidence either did not belong to the defendant or was lawfully seized. On April 1, 2007, two months before the start of trial, The Cape Cod Times, a regional paper serving the Cape and Islands, published an article entitled, “Toolan Claiming Insanity,” which reported that Toolan would be using an “insanity defense” (emphasis added), and discussed the low success rate of this “defense.” The article quoted a Falmouth defense attorney, not involved in the defendant’s case, who stated: “Toolan appeared lucid enough to get himself to Nantucket to commit the crime .... Nothing I’ve read about [the defendant] so far makes me think he had a history of mental *457illness, but who knows? ... He seemed to be highly functioning in society from what I read. But that’s not conclusive.”
On the eve of jury empanelment, which began June 4, 2007, news stories previewed the upcoming trial. From The Nantucket Independent published Wednesday, May 30, 2007;3 The Inquirer and Mirror published May 31, 2007;4 *and The Cape Cod Times published Sunday, June 3, 2007,5 readers could learn once again of the reported facts of the case, including the allegation that the defendant held the victim prisoner in his New York apartment days before killing her. All three stories previewed the upcoming trial, and all three stated the defendant would argue insanity. The two Nantucket papers included commentary on the “insanity defense,” which they described as a “difficult” or “uncommon” strategy that “rarely succeeds.” The two Nantucket papers also suggested a burden on the defense to “prove” some component of diminished capacity, such as intoxication or mental illness. Articles continued to appear throughout the trial, including the three days of jury empanelment, June 4-6.
During the jury empanelment, some members of the jury ve-nire explained to the judge their many connections to the victim or her family.6 In all, forty-three members of the venire, or *458twenty-five per cent of the total over three days, stated in individual voir dire that they knew the family directly or indirectly.7
Another sixty-seven jurors, or thirty-eight per cent, did not report knowing the victim or her family but said they knew witnesses who were scheduled to testify. These relationships ranged from intimate •— several jurors were related to witnesses by blood or marriage — to casual. Some members of the venire knew the medical examiner as their personal physician. Many knew police witnesses. Others knew the victim’s landlady as a close family friend or through her work on the board of a local organization. Still others knew witnesses as neighbors, employers, coworkers, customers, and friends. Even though the judge did not systematically question the venire covering media exposure or knowledge of the case, fifty prospective jurors (twenty-nine per cent) reported to the judge that they had talked about the case prior to trial, while seventy-four (forty-two per cent) mentioned they had seen media coverage of the case.
b. Empanelment procedure. During jury empanelment, with some variations described below, the judge relied on the following general procedure to examine jurors on each of the three days. He first spoke collectively to the members of each day’s group of prospective jurors on the topics of the “statutory questions,”8 the Commonwealth’s burden of proof, and anticipated *459witnesses. Next, he questioned individually those jurors who had raised their hands in response to the matters to which the judge had referred his collective remarks.9 In doing so, the judge followed no set pattern but addressed the particular concerns that jurors indicated had led them to raise their hands.10 A total of 120 out of 175 jurors were excused as a result of this questioning. Each of the fifty-five remaining jurors was questioned individually in what amounted to a second phase of the process. The individual questioning in this second phase was more uniform, and focused on the issue of lack of criminal responsibility;11 the juror’s exposure to violent crime or domestic violence, and to mental illness; the juror’s opinion on psychiatrists and lawyers; the juror’s view of the defendant’s right to remain silent at trial; and whether the juror had visited the cottage where the victim died. Jurors were not asked about their previous discussions regarding or knowledge of the case, their exposure to pretrial publicity, or their acquaintance with potential witnesses.
Moving from the general to the particular, the record reflects that on the first day of jury empanelment, the judge briefly summarized for the prospective jurors the Commonwealth’s allegations and theories of the case. He outlined the presumption of innocence and the Commonwealth’s burden to prove guilt beyond a reasonable doubt. He urged the venire to “examin[e] [their] conscience [s] relative to whether or not [they could] remain impartial, open minded, and indifferent.” He stated that the case had “received an inordinate amount of publicity” and then outlined what he termed “area[s] of the court’s concern,” including whether publicity had affected their ability to be *460impartial; whether they had any interest in the case or had expressed or formed an opinion about the case; whether they were sensible of any bias or prejudice that they might have with respect to the case. He introduced the counsel for the Commonwealth and for the defense. The judge then read the names of about 120 witnesses (filling about ten pages of transcript). Up to this point, the judge had asked the venire no questions, but simply instructed the group that they should consider areas of his concern.
At the conclusion of this reading, the judge asked for a collective show of hands “from all of you who feel that you cannot sit on this case for any of the reasons that I’ve stated.” Without recording which jurors raised their hands, the judge instructed them to remember whether they had done so.12 A lunch break followed; it lasted over an hour for the venire.
After lunch, the defendant’s attorney formally objected to the judge’s method of voir dire and suggested that the judge ask for a show of hands at the end of each separate question and then record the numbers of the jurors with raised hands. The judge declined at that point to alter his method. On returning from their lunch break, those jurors who remembered that they had raised their hands were called individually to speak with the judge in the presence of both counsel. Some did not appear to understand the self-identification procedure or were unable to remember their affirmative answers to the “areas of the court’s concern.”
Six individuals who eventually served as jurors on the case did not raise their hands in response to the judge’s initial general questions — or did not remember that they had raised their hands — and so were never individually questioned on issues other than lack of criminal responsibility and related topics during the second phase of the empanelment process. The two other seated jurors drawn from the first day’s venire did raise their hands and were questioned following the judge’s general *461questions. However, the judge did not ask the first whether she had been exposed to publicity about, or had knowledge of, the case. The second was asked whether she had read about the case and whether she remembered when it happened, and she acknowledged that she had and she did, but the judge did not ask her any further questions, including whether she had formed any opinion of the case.
On the morning of the second day of empanelment, the judge addressed the second group of prospective jurors much as he had the first group the day before, although midway through his remarks, the judge briefly restated the “area[s] of concern” of the statutory questions. He then read the list of witnesses, concluded his remarks, and commenced individual questioning of every member of the second venire but one, apparently because they all raised their hands.13
On the morning of the third day, the judge addressed the final group of prospective jurors. This time, the judge restated the areas of concern of the statutory questions at the very end of his general introductory remarks, asked for a show of hands immediately following the restatement, and directly began individual questioning. In the afternoon, the judge individually questioned the nineteen prospective jurors retained from the second day on lack of criminal responsibility issues, and then repeated the process for the twelve jurors retained from that morning. After completing these two sets of individual voir dire, the judge denied the defendant’s motion for change of venue, explaining that he had “compiled a list of [forty-six] indifferent, impartial jurors to sit in this case.”14 On the fourth day, after the parties’ exercise of peremptory challenges,16 sixteen members of the ve-nire were seated as jurors.
In summary, six of the seated jurors, all from the first day’s venire of prospective jurors, were never questioned by the judge *462concerning exposure to publicity, other knowledge of the case or of witnesses, or any similar issue. Of these six, two were ultimately designated as alternates, but the other four participated in the verdicts.
c. Presumptive prejudice. Article 12 of the Massachusetts Declaration of Rights and the Sixth Amendment to the United States Constitution guarantee the right of a criminal defendant to a trial by an impartial jury. Commonwealth v. Susi, 394 Mass. 784, 786 (1985) (“The failure to grant a defendant a fair hearing before an impartial jury violates even minimal standards of due process”). See Irvin v. Dowd, 366 U.S. 717, 722 (1961); Commonwealth v. Guisti, 434 Mass. 245, 251 (2001). In reviewing a defendant’s claim that his right to trial by an impartial jury was violated as a result of prejudice from extensive pretrial publicity or settled community opinion, we examine first whether a change of venue was required because the jury were presumptively prejudiced against him. See Skilling v. United States, 130 S. Ct. 2896, 2912 (2010) (Skilling); Commonwealth v. Morales, 440 Mass. 536, 540-542 (2003) (Morales). If the jury were not presumptively prejudiced, we next examine whether the defendant has shown actual juror prejudice. Skilling, supra. Morales, supra at 542.
The defendant claims that in the circumstances of the case, the judge abused his discretion and therefore committed error of law in denying a change of venue. He argues that the case met the criteria set out by the United States Supreme Court and this court for presumptive prejudice, stating that publicity was extensive and sensational and that the venue was an exceedingly small, nonurban community. He also argues that the high percentage of jurors who admitted to disqualifying prejudice cast doubt on the avowals of impartiality by remaining jurors.
A trial judge may order a change of venue if “there exists in the community where the prosecution is pending so great a prejudice against the defendant that he may not there obtain a fair and impartial trial.” Mass. R. Crim. R 37 (b) (1), 378 Mass. 914 (1979). A change of venue should be ordered only “with great caution and only after a solid foundation of fact has been first established.” Commonwealth v. McCowen, 458 Mass. 461, 476 (2010), quoting Commonwealth v. Clark, 432 Mass. 1, 6 *463(2000). A trial judge has substantial discretion in deciding whether to allow a motion for a change of venue; we review for abuse of discretion. See Commonwealth v. McCowen, supra at 475.
Prejudice against the defendant sufficient to preclude a fair and impartial trial may exist because the entire jury pool is tainted by exposure to pretrial publicity. See, e.g., Irvin v. Dowd, 366 U.S. at 726-728. In such circumstances, the venire is considered presumptively prejudiced, regardless of the details of the voir dire process, see Rideau v. Louisiana, 373 U.S. 723, 724-726 (1963) (Rideau), and even if individual members of the jury expressly assert their belief that they can be “fair and impartial.” See Irvin v. Dowd, supra at 727-728.16 However, presumptive prejudice exists only in truly extraordinary circumstances.17 Commonwealth v. Leahy, 445 Mass. 481, 494-495 (2005).
“[P]retrial publicity — even pervasive, adverse publicity — does not inevitably lead to an unfair trial.” Skilling, 130 S. Ct. at 2916, quoting Nebraska Press Ass’n v. Stuart, 427 U.S. 539, 554 (1976). In assessing whether such publicity and resulting local prejudice precludes a fair trial, the United States Supreme Court in Skilling looked to the influence, if any, of the media on the trial; the size of the community; the content of the news stories;18 the length of time between the peak media coverage and the trial; and any evidence from the verdict itself, such as *464the jury’s decision to acquit the defendant of any of the charges. Id. at 2913-2916. This court has likewise identified the nature of the publicity (whether “extensive and sensational”) as a highly significant factor.19 Morales, 440 Mass. at 540. However, we have attached primary importance to the actual ability of the trial judge to empanel jurors who appear impartial. Id. at 541. See Commonwealth v. Angiulo, 415 Mass. 502, 514-516 (1993).20
The question is a close one. The small size of the Nantucket community weighs in favor of finding local prejudice. In Skilling, 130 S. Ct. at 2915, the Supreme Court contrasted Houston’s large pool of 4.5 million potential jurors against the 150,000 residents of the community in Rideau. See Commonwealth v. Leahy, 445 Mass. at 493-494 (Plymouth County, part of metropolitan Boston, “a far cry” from rural Indiana). Nantucket, with just over 10,000 permanent residents, is smaller still. The extensive links among the victim’s family, members of the over-all jury venire, and trial witnesses demonstrate the network of social relations connecting this community, of which the victim was a valued member, and to which the defendant was an outsider.21
*465Likewise, the media coverage of the case was extensive, at times emotional, and prejudicial to the defendant’s anticipated theory of defense. Understandably enough, publicity portrayed the death of a charismatic woman as a tragedy that had brought grief to her sympathetic family and friends. Portrayals of the defendant emphasized his troubled history with women, his difficulties in holding a job, his erratic criminality, and his alcohol problems while simultaneously suggesting he was unlikely to succeed on his theory that he was not guilty by reason of insanity. Moreover, the media’s discussion of the defendant’s theory of defense at trial could have left readers with the incorrect impression that the defendant carried the burden of proving his lack of criminal responsibility at trial, and a heavy burden at that.
Nonetheless, other factors cause us to conclude the jury pool was not pervasively biased by media coverage. The defendant has presented no evidence of a raucous or carnival atmosphere at trial. See Skilling, supra at 2914. Contrast Sheppard v. Maxwell, 384 U.S. 333, 355, 358 (1966) (“bedlam reigned at the courthouse during the trial” due to presence of “newsmen,” creating pervasive “carnival atmosphere”). Moreover, a lapse of almost three years between the victim’s death and the trial was likely to have blunted the impact of initial media coverage. See Patton v. Yount, 467 U.S. 1025, 1032 (1984) (no manifest error in finding of jury impartiality where jury selection occurred four years after crime, when “prejudicial publicity was greatly diminished and community sentiment had softened”); Morales, 440 Mass. at 540-541 (intensity of media coverage dissipated in fourteen months between death and trial). Contrast Irvin v. Dowd, 366 U.S. at 725, 727-728 (summarized in note 16, supra).
By comparison with Irvin and Rideau (see note 16, supra), the content of the media stories in this case was not as prejudicial to the defendant. While the news media tended to paint the defendant as a “Jekyll and Hyde” character — i.e., a seemingly pleasant man with “a dark history of erratic behavior” and a serious alcohol problem — pretrial publicity left open questions of the relationship between or among his mental problems, alcohol, and the victim’s death. The likely impact on potential jurors differed markedly from the impact in Rideau, in which the televised confession directly contradicted Rideau’s not guilty *466plea and “in a very real sense was Rideau’s trial” (emphasis in original). Rideau, 373 U.S. at 726. In this case, as in Skilling, 130 S. Ct. at 2916, the pretrial publicity did not amount to the “smoking gun variety” certain to invite prejudgment of culpability.
In the final analysis, we conclude that the defendant has not shown that, as a result of pretrial publicity or otherwise, the community’s prejudgment of the case was substantial enough that empanelling impartial jurors would not be possible. See Morales, 440 Mass. at 541-542. His claim of presumptive prejudice requiring a change of venue fails.22 After remand, on the defendant’s motion, the trial judge would have the discretion to grant a change of venue should the judge find the defendant cannot obtain a fair trial on Nantucket. See note 17, supra.
d. Actual prejudice. We turn, therefore, to whether, in the totality of the circumstances, the publicity and the particular venue at issue — Nantucket — deprived the defendant of his right to a fair trial. The defendant argues that even if an impartial jury theoretically could have been selected on Nantucket, the inadequacy of the jury selection process in his case necessitates reversal because it produced a biased jury.
Where, because of pretrial publicity or local community issues, the risk of juror bias is particularly acute, the United States Supreme Court has placed a correspondingly high emphasis on the need for an adequate voir dire. See, e.g., Skilling, 130 S. Ct. at 2917 (concluding that “widespread community impact” required careful inquiry into jurors’ connections to case).23
This court also has placed a strong emphasis on the voir dire *467process where the risk of prejudice to the defendant from pretrial publicity has been especially significant.24 See Commonwealth v. McCowen, 458 Mass. at 476 (risks from “substantial pretrial publicity” obviated by individual voir dire); Commonwealth v. Leahy, 445 Mass. at 495-496 & n.13 (court relied on judge’s “credibility determinations” as to jurors’ assertions of impartiality in individual voir dire); Morales, 440 Mass. at 538, 542-543 (judge conducted very careful individual voir dire to select jurors “unswayed by any media publicity”; conviction affirmed); Commonwealth v. Clark, 432 Mass. at 5-6 (despite “massive publicity,” denial of change of venue not error where each juror *468questioned in individual voir dire on impartiality). A juror need not be “totally ignorant of the facts and issues involved” so long as “the juror can lay aside his impression or opinion and render a verdict based on the evidence presented in court.” Commonwealth v. Leahy, supra at 494, quoting Irvin v. Dowd, 366 U.S. at 722-723. Where potential jurors have had pretrial exposure to a case, individual voir dire is the means by which a judge can distinguish between mere exposure and bias. See, e.g., Commonwealth v. McCowen, supra. See also Commonwealth v. Stroyny, 435 Mass. 635, 638-639 (2002) (jurors who indicated they had any knowledge of case from pretrial publicity individually questioned concerning impartiality); Commonwealth v. Clark, 432 Mass. at 6 (change of venue correctly denied where all seated jurors individually told judge they could be fair and impartial and all were determined to be indifferent).
In the present case, we conclude that in the totality of the circumstances, the defendant was denied his right to a fair trial. For reasons previously discussed, we have not found presumptive prejudice in the jury pool. Nevertheless, the circumstances were such that for any particular prospective juror, the risk of bias was high. Pretrial publicity was extensive, and it was at times prejudicial to the defendant’s anticipated theory of defense.25 In addition, in the small, socially interconnected community of Nantucket, in which many prospective jurors knew the victim, her family, or witnesses, the influence of mass media coverage overlapped with, and was reinforced by, informal interactions and conversations. Some prospective jurors had second-hand or even direct factual knowledge of the case prior to trial; others had developed opinions through conversations with family members, friends, and acquaintances. In this context, as in Skilling, McCowen, Morales, and other cases involving high-profile prosecutions, it was essential that the judge carefully assess whether each juror could be impartial.
*469The judge failed to meet this obligation. As demonstrated by his repeated statements that the case had attracted considerable publicity, he apparently concluded, and the Commonwealth essentially concedes, that exposure to potentially prejudicial material created a risk of bias among members of the jury pool. But even though he recognized the possibility of such exposure to “issues extraneous to the case,” G. L. c. 234, § 28, the judge did not conduct his examination in a way that would have allowed him to make a sound determination as to whether each juror was impartial despite the exposure. Instead, he shifted the burden of assessing impartiality from himself to individual members of the venire.26 Jurors’ answers to individual voir dire questions on the defense of lack of criminal responsibility provided some additional insight into their willingness to return not guilty verdicts, but did not directly address the impact of publicity and local community attitudes and their opinion on them.
Because the voir dire was insufficient to support a contrary conclusion, it is appropriate to assume in the context of this case that jurors who did not state otherwise were exposed to negative pretrial publicity. See Commonwealth v. Crehan, 345 Mass. 609, 613, 615 (1963) (where prejudicial newspaper articles about defendants were published during trial and trial judge declined to poll jurors on exposure to, or impact of, articles, court assumed jurors had read them, and ordered new trial).27 Contrast Commonwealth v. Clark, 432 Mass. at 18 (Crehan distinguished where judge questioned jurors individually about influence from extraneous sources). Given the nature of the publicity, especially in combination with the circumstance of the trial taking place within the small island community of Nantucket and the victim’s connection to it, we simply cannot say with confidence that the defendant was tried by an impartial *470adjudicator and found guilty based solely on evidence presented at trial. See Commonwealth v. Susi, 394 Mass. 784, 787 (1985), quoting Commonwealth v. Stanley, 363 Mass. 102, 104 (1973) (defendant “entitled to have the case decided only upon the evidence that was introduced at the trial”). See also Patterson v. Colorado ex rel. Attorney Gen., 205 U.S. 454, 462 (1907) (Holmes, J.) (“The theory of our [trial] system is that the conclusions to be reached in a case will be induced only by evidence and argument in open court, and not by any outside influence, whether of private talk or public print”).
This error admits of no easy remedy. We do not think it would be practical or helpful to remand for a postconviction evidentiary hearing. The jurors themselves may no longer know whether, prior to hearing the evidence at trial, they were biased in favor of finding the defendant guilty on the basis of pretrial media or community opinion. Contrast Commonwealth v. Johnson, 426 Mass. 617, 620, 625-628 (1998) (juror asked simple question of fact in postconviction evidentiary hearing, i.e., whether her husband had, to her knowledge, attended victim’s wake). Paradoxically, the very nature of the error makes it impossible to discern whether seated jurors held subtle biases against the defendant — biases that might have become apparent in individual questioning during empanelment on the sources of their knowledge of the case, and any opinions formed as a result. Accordingly, we reverse the defendant’s convictions and remand for a new trial.
3. Comment at trial on Miranda rights. The defendant contests the introduction of evidence of his statements to the police that he understood his Miranda rights, his deliberations about his rights, his attempt to call someone for “advice,” and his refusal to waive his Miranda rights and make a statement. He argues that the Commonwealth’s evidence improperly exploited his exercise of his constitutional right to silence to strike at his defense of lack of criminal responsibility by showing that he was capable of understanding and careful deliberation rather than acting impulsively. Because the issue will likely recur on remand, we address it.
A defendant’s exercise of his right to remain silent or to consult counsel may not be used as substantive evidence of guilt or as evidence negating a claim of insanity. Wainwright v. Green*471field, 474 U.S. 284, 295 (1986). Commonwealth v. Mahdi, 388 Mass. 679, 694-695 & n.15 (1983). See Doyle v. Ohio, 426 U.S. 610, 613-619 (1976). “Miranda warnings contain an ‘implicit assurance that a defendant’s silence after such warnings will carry no penalty,’ and due process requires that, when in the hands of the police, a defendant must be able to ‘invoke core constitutional rights without fear of making implied or adoptive admissions.’ ” Commonwealth v. Beneche, 458 Mass. 61, 73 (2010), quoting Commonwealth v. Peixoto, 430 Mass. 654, 657, 658-659 (2000). Mention of Miranda warnings is not, however, forbidden at trial: under the Commonwealth’s “humane practice” rule, where the voluntariness of a statement is a live issue at trial, the jury may hear evidence that a defendant was informed of and understood his Miranda rights. See, e.g., Commonwealth v. Parker, 412 Mass. 353, 357-358 & n.11 (1992).
The focus of the defendant’s argument is testimony of a State trooper and comments by the prosecutor. In particular, the Massachusetts State trooper who questioned the defendant in Rhode Island testified that after the trooper advised the defendant of his rights, the defendant responded that he understood his rights but said he did not want to sign a Miranda waiver. The trooper informed the defendant that the police wished to talk to him regarding an “incident,” and in response to questions from the defendant, said the incident had occurred that day, and involved the victim. The trooper told the defendant several times he would not answer the defendant’s questions unless the defendant signed a waiver form. The defendant claimed he could not have had any involvement in any incident with the victim that day (October 25), as he had not seen her since Friday, October 22. The trooper also testified the defendant attempted to place a telephone call “for advice” to a person the State trooper identified by name but did not describe as a lawyer.
The prosecutor’s challenged comment was in his opening statement. He referred to anticipated testimony by State troopers as to their observations of the defendant while he was deciding whether to exercise his Miranda rights. Defense counsel objected to this statement when made, and the judge provided a curative instruction.28
*472It was permissible for the prosecutor, without reference to Miranda rights, to introduce and comment on evidence to the effect that the defendant’s actions upon his arrest, including his apparent attempt to use his interview with the police to acquire information, showed his sanity. As the Supreme Court noted in Wainwright v. Greenfield, 474 U.S. at 295, “the State’s legitimate interest in proving that the defendant’s behavior appeared to be rational at the time of his arrest could have been served by carefully framed questions that avoided any mention of the defendant’s exercise of his constitutional rights to remain silent and to consult counsel.” The Commonwealth also was entitled to present evidence that the defendant was informed of his Miranda rights where, as here, the defendant challenged the voluntariness of statements made in custody. See Commonwealth v. Parker, 412 Mass. at 357-358 & n.11. Nonetheless, the references in the State trooper’s testimony and the prosecutor’s opening statement to the defendant’s constitutional rights come dangerously close to an improper suggestion that the defendant was manipulating his constitutional rights to his own advantage.
On remand, the Commonwealth should exercise care to avoid using the defendant’s exercise of his Miranda rights against him by suggesting that his invocations of or deliberations on these rights demonstrated his criminal responsibility. The Commonwealth may elicit testimony that the police administered Miranda warnings, and that the defendant stated that he understood them. The defendant’s unsolicited statements and questions to police about the nature of the police investigation and about the victim may also be introduced. However, unless the defendant opens the door, it is not proper for the Commonwealth to elicit testimony referencing the defendant’s deliberations over whether to exercise his Miranda rights, see Commonwealth v. Mahdi, 388 Mass,. at 697-698, or his decision to contact an *473attorney and efforts to do so. See Commonwealth v. DePace, 433 Mass. 379, 384 (2001), S.C., 442 Mass. 739 (2004), cert. denied, 544 U.S. 980 (2005) (request for counsel only rarely admissible, e.g., to explain abrupt end to interview with police).
4. Instruction on mental defect. Where there is evidence at a murder trial that the defendant had a mental disease or defect and consumed drugs or alcohol, the jury must be properly instructed on the intersection of voluntary intoxication and mental disease or defect. See Commonwealth v. DiPadova, ante 424, 430-433, 439-440 (2011) (setting forth appropriate instructions). See also Commonwealth v. Berry, 457 Mass. 602, 617-618 & n.9 (2010). The defendant claims the instructions as given were erroneous in light of Berry, which was decided after the trial in this case. The Commonwealth disagrees, and claims that Berry does not apply in this case because, inter alia, the defendant was aware of the effect alcohol had on him.29 The Commonwealth’s argument erroneously appears to assume that the defendant’s level of awareness or knowledge of the impact of alcohol on his mental state was an uncontested fact rather than a question for the jury. In any event, at retrial, the judge should consider the application of DiPadova and Berry to the case in light of the particular evidence introduced, and instruct the jury accordingly.
5. Conclusion. For the foregoing reasons, the defendant’s convictions are reversed, the verdicts set aside, and the case is remanded for a new trial in accordance with this opinion.
So ordered.
The defendant also was found guilty of assault and battery by means of a dangerous weapon of the same victim.
The town and county of Nantucket are coterminous. See, e.g., G. L. c. 57, § 4. The jury pool consisted entirely of Nantucket residents. See G. L. c. 234A, §§ 1-3.
The Nantucket Independent ran a front page story on May 30, 2007, anticipating, as stated in the headline, Nantucket’s “first murder trial in 24 years.” Photographs of the prosecutor and defense attorney were visible above the fold; below the fold there was a photograph of the defendant wearing a suit and handcuffs and flanked by police officers. The article included biographical facts about the prosecutor, defense attorney, and judge, a detailed chronology of the events leading up to and including the murder, and a diagram of the empty court room. The article quoted a former local assistant district attorney stating that the defense would present a “parade” of medical experts in support of his insanity plea and described the views of three other attorneys and a forensic psychiatrist on the insanity defense.
On May 31, 2007, The Inquirer and Mirror’s front page story, “Toolan goes on trial Monday for 2004 murder of Beth Lochtefeld,” reviewed the defendant’s past, the murder, and the feelings of one of the victim’s brothers.
As will be discussed infra, the empanelment process included individual questioning of the prospective jurors on lack of criminal responsibility, but *458there was no systematic individualized inquiry into jurors’ exposure to publicity about, or community discussion of attitudes relating to, the case. Members of the jury who supplied the information described in the text, about their connections to the victim and her family or to other witnesses, did so when they spoke to the judge because they had raised their hands in response to a group of questions he posed to the venire as a whole.
On June 3, 2007, the day before the trial began, the Sunday edition of The Cape Cod Times featured the victim’s and defendant’s photographs on the front page with an article focusing on the victim’s life and murder. The article included a photograph of the friend who introduced the victim and the defendant to one another, wearing an anguished expression at the defendant’s arraignment, and a photograph of the cottage where the victim’s murder took place.
Ten knew the victim directly; another’s daughter was friends with the victim; still another worked for a friend of the victim. Seventeen more did not know the victim but knew her brother or his wife. Ten more potential jurors, exclusive of those already mentioned, knew one or both of the victim’s parents. Another four had less direct connections to the family.
As a matter of statute, on the motion of either party, every juror must be examined “to learn whether he [1] is related to either party or has any interest in the case, or [2] has expressed or formed an opinion, or [3] is sensible of any bias or prejudice, therein.” G. L. c. 234, § 28. In a criminal case, the examination must also include “questions” on “[4] whether such juror understands that a defendant is presumed innocent until proven guilty, [5] that the commonwealth has the burden of proving guilt beyond a reasonable doubt, and [6] that the defendant need not present evidence in his behalf.” Id. For brevity, we refer to these questions as the “statutory questions.”
Of the entire pool of 175 jurors, 156 were questioned individually after raising their hands, including sixty-one out of seventy-seven on the first day, sixty-four out of sixty-five on the second day, and thirty-one out of thirty-three on the third day.
Topics included responses to the statutory questions as well as juror requests for hardship exemptions, for example for travel plans, family commitments, or medical issues.
Individual voir dire on the lack of criminal responsibility was required here. See Commonwealth v. Seguin, 421 Mass. 243, 249 (1995), cert. denied, 516 U.S. 1180 (1996) (where defendant raises lack of criminal responsibility, jurors must be asked individually whether they hold opinion preventing them from returning verdict of not guilty by reason of insanity if Commonwealth fails in its burden to prove criminal responsibility).
As noted in Commonwealth v. Shea, ante 163, 169 (2011), we discourage use of a protocol that relies on jurors making a “mental note” of their affirmative answers to questions asked of the venire. It is far preferable for trial judges to instruct jurors to raise their hands or otherwise promptly and visibly indicate affirmative answers in response to each question individually. See Commonwealth v. Tatro, 42 Mass. App. Ct. 918, 920 (1997).
During the afternoon of the second day, the judge conducted the individual voir dire of the twenty-four prospective jurors who had been retained from the first day on lack of criminal responsibility and related issues.
The defendant filed a petition for relief under G. L. c. 211, § 3, and interlocutory relief under Mass. R. Grim. R 15, as appearing in 422 Mass. 1501 (1996), seeking review of this decision. A single justice of this court denied the petition.
One juror was excused for cause at this stage.
In Irvin v. Dowd, 366 U.S. 717 (1961), newspapers delivered regularly to ninety-five per cent of dwellings in the county from which the jury were drawn announced prior to trial that the defendant had confessed to six murders, and ninety per cent of those jurors examined on the point in voir dire “entertained some opinion” as to the defendant’s guilt while two-thirds of the seated jury members expressed an opinion in voir dire that the defendant was guilty. Id. at 725, 727-728. The facts in Rideau v. Louisiana, 373 U.S. 723 (1963) (Rideau), also were extraordinary: a filmed custodial “interview” in which the jailed defendant confessed to murder was rebroadcast on local television three nights in a row less than two months prior to trial. Id. at 724. Id. at 729 (Clark, J., dissenting).
Although trial judges can and do exercise their discretion to transfer a case to another county on the basis of pretrial publicity in appropriate circumstances, see, e.g., Commonwealth v. Gaynor, 443 Mass. 245, 258-259 (2005), the defendant points to no case in which this court has overturned a criminal conviction on the basis of presumptive bias in the jury pool, and we are aware of no such case.
In Skilling v. United States, 130 S. Ct. 2896, 2916 (2010) (Skilling), the United States Supreme Court differentiated between news stories that are “not *464kind” to the defendant but contain no confession or blatantly prejudicial information and “smoking-gun” publicity such as the “dramatically staged admission of guilt” in Rideau, 373 U.S. at 724; the Court concluded the former may not invite prejudgment of culpability.
Media coverage that is inflammatory or sensational as opposed to factual better supports a claim for change of venue. Commonwealth v. Morales, 440 Mass. 536, 540 (2003). Factual coverage may include, but is not limited to, summaries of the incident, descriptions of the charges against the defendant, and information about the defendant’s arrest, while sensational coverage may contain more emotional, sensational material. Id.
We do not agree with the defendant that the sheer number of jurors disqualified for prejudice casts doubt on the impartiality of the remaining jurors. In previous cases, this court has held that excusing up to forty-two per cent of the venire due to prejudice from pretrial publicity was insufficient to show that it was impossible to empanel an impartial jury. See Commonwealth v. Clark, 432 Mass. 1, 6 (2000) (thirty-five per cent of venire); Commonwealth v. Angiulo, 415 Mass. 502, 515 (1993) (forty-two per cent of venire). See also Commonwealth v. Morales, 440 Mass. at 541 & n.1 (twenty-five per cent of venire). In this case, about twenty-six per cent of the venire was excused at least in part due to prejudice from pretrial publicity.
Strikingly, only one person questioned in individual voir dire mentioned any connection to the defendant. Connection, in this case, did not equate with positive bias: he was friends with the defendant’s former girl friend, and “it sure seem[ed] to [him] [the defendant] was guilty.”
The defendant used only fourteen of his sixteen peremptory challenges. Generally, a defendant’s failure to exhaust his peremptory challenges weighs against finding that prejudice necessitated a change of venue. See Commonwealth v. Morales, 440 Mass. at 543; Delle Chiaie v. Commonwealth, 367 Mass. 527, 532 (1975). Contrast Irvin v. Dowd, 366 U.S. at 724 (noting that defendant exhausted peremptory challenges in finding jury pool tainted). In this case, however, we are able to draw no conclusions; because the judge identified forty-six impartial jurors before the exercise of any peremptory challenges, the defendant knew as much about the jurors farther down the list as he did about the seated jurors. He may have been satisfied with the jury after using only fourteen challenges, but alternately, he may have concluded that the seated jurors were better than or no worse than the remaining lot.
The defendant in Skilling had been a high-level executive at an energy company, Enron Corporation, that collapsed spectacularly in 2001; Skilling and others were prosecuted after an “investigation uncovered an elaborate *467conspiracy to prop up Enron’s short-run stock prices by overstating the company’s financial well-being.” Skilling, 130 S. Ct. at 2907. Thousands of Enron employees in Houston lost their jobs and their retirement savings. Id. at 2942 (Sotomayor, J., concurring in part and dissenting in part). Against a background of extensive negative publicity, the Supreme Court in Skilling assessed actual prejudice to the defendant almost entirely in terms of whether the voir dire of prospective jurors was adequate. Id. at 2917-2925. A majority of the Court concluded that the voir dire conducted by the trial judge, which included a lengthy questionnaire sent to each prospective juror before trial and follow-up questions asked by the judge of each individual prospective juror during empanelment, was sufficient to provide a fair and impartial jury in accord with the guarantee of the Sixth Amendment to the United States Constitution. See id. at 2920, 2923. However, even with the judge’s use of the questionnaire and individual juror questioning, four of the Justices expressed serious concern about the adequacy of the voir dire. See id. at 2941 (Alito, J., concurring in part and concurring in the judgment); 2956-2963 (Sotomayor, J., concurring in part and dissenting in part).
Statute and court rule alike require that where “issues extraneous to the case” such as “community attitudes [or] possible exposure to potentially prejudicial material” mean jurors “may not stand indifferent,” jurors must be examined on the issue “individually and outside the presence of other [jurors].’’ G. L. c. 234, § 28. See Mass. R. Crim. P. 20 (b) (2), 378 Mass. 889 (1979). “An extraneous issue is one that goes beyond the record and raises a serious question of possible prejudice.” Commonwealth v. Kater, 432 Mass. 404, 414 (2000). A trial judge has discretion in determining whether a preliminary foundation has been laid such that an individual voir dire becomes necessary. Commonwealth v. Shelley, 381 Mass. 340, 352 (1980). But once a judge determines that such a foundation exists, G. L. c. 234, § 28, “impose[s] a duty on the judge to examine jurors fully with respect to possible bias or prejudice.” Id., quoting Commonwealth v. Dickerson, 372 Mass. 783, 793 (1977). This statutory scheme aligns with the duties of a trial judge to protect a defendant’s constitutional right to an impartial jury, free from “interest, bias and prejudice.” Commonwealth v. Beneficial Fin. Co., 360 Mass. 188, 295 (1971), cert. denied sub nom. Farrell v. Massachusetts, 407 U.S. 910, and sub nom. Beneficial Fin. Co. v. Massachusetts, 407 U.S. 914 (1972).
Media coverage also continued during trial, but the judge questioned the sitting jurors each trial day about exposure to publicity and admonished them not to read the newspapers or discuss the case. Jurors did not indicate to the judge that they were exposed to media about the case during trial. Absent evidence to the contrary, jurors are presumed to follow the judge’s instructions. See Commonwealth v. Lynch, 439 Mass. 532, 544, cert. denied, 540 U.S. 1059 (2003).
With respect to four jurors who participated in the verdict, the judge relied entirely on their silence in response to his statement that it was “an area of the court’s concern” if pretrial media coverage had “affected [their] ability to be open minded, impartial[,] and indifferent.” Contrast G. L. c. 234, § 28; Commonwealth v. Leahy, 445 Mass. at 495-496 & n.13.
Defense counsel in this case requested that the judge ask prospective jurors to complete a case-specific questionnaire that included questions concerning exposure to publicity about the case and their knowledge of it. There is nothing in the record that indicates the judge granted the request or that such a questionnaire was used.
In his closing, in the context of an argument that the defendant’s actions *472after the killing showed a high level of planning and impulse control, the prosecutor suggested the defendant was using the interview with the police in an attempt to gain information. The prosecutor noted that the defendant had pressed for more information while the police had only asked, “Will you sign the waiver?” The prosecutor then asked the jury, “[I]f he’s so intoxicated, how is he able to conduct what I would suggest to you requires a rather high level of thinking in order to interview the interviewers?” The defendant does not directly challenge this argument on appeal.
After this case was briefed and argued in this court, we decided Commonwealth v. DiPadova, ante 424 (2011). Like Commonwealth v. Berry, 457 Mass 602, 613-618 (2010), DiPadova addressed the intersection of intoxication and mental illness. It slightly modified the instructions required where there is evidence that a defendant had a mental disease or defect and voluntarily consumed drugs or alcohol. See Commonwealth v. DiPadova, supra at 439-440.